GPU-accelerated adjoint algorithmic differentiation
NASA Astrophysics Data System (ADS)
Gremse, Felix; Höfter, Andreas; Razik, Lukas; Kiessling, Fabian; Naumann, Uwe
2016-03-01
Many scientific problems such as classifier training or medical image reconstruction can be expressed as minimization of differentiable real-valued cost functions and solved with iterative gradient-based methods. Adjoint algorithmic differentiation (AAD) enables automated computation of gradients of such cost functions implemented as computer programs. To backpropagate adjoint derivatives, excessive memory is potentially required to store the intermediate partial derivatives on a dedicated data structure, referred to as the "tape". Parallelization is difficult because threads need to synchronize their accesses during taping and backpropagation. This situation is aggravated for many-core architectures, such as Graphics Processing Units (GPUs), because of the large number of light-weight threads and the limited memory size in general as well as per thread. We show how these limitations can be mediated if the cost function is expressed using GPU-accelerated vector and matrix operations which are recognized as intrinsic functions by our AAD software. We compare this approach with naive and vectorized implementations for CPUs. We use four increasingly complex cost functions to evaluate the performance with respect to memory consumption and gradient computation times. Using vectorization, CPU and GPU memory consumption could be substantially reduced compared to the naive reference implementation, in some cases even by an order of complexity. The vectorization allowed usage of optimized parallel libraries during forward and reverse passes which resulted in high speedups for the vectorized CPU version compared to the naive reference implementation. The GPU version achieved an additional speedup of 7.5 ± 4.4, showing that the processing power of GPUs can be utilized for AAD using this concept. Furthermore, we show how this software can be systematically extended for more complex problems such as nonlinear absorption reconstruction for fluorescence-mediated tomography.
GPU-Accelerated Adjoint Algorithmic Differentiation
Gremse, Felix; Höfter, Andreas; Razik, Lukas; Kiessling, Fabian; Naumann, Uwe
2015-01-01
Many scientific problems such as classifier training or medical image reconstruction can be expressed as minimization of differentiable real-valued cost functions and solved with iterative gradient-based methods. Adjoint algorithmic differentiation (AAD) enables automated computation of gradients of such cost functions implemented as computer programs. To backpropagate adjoint derivatives, excessive memory is potentially required to store the intermediate partial derivatives on a dedicated data structure, referred to as the “tape”. Parallelization is difficult because threads need to synchronize their accesses during taping and backpropagation. This situation is aggravated for many-core architectures, such as Graphics Processing Units (GPUs), because of the large number of light-weight threads and the limited memory size in general as well as per thread. We show how these limitations can be mediated if the cost function is expressed using GPU-accelerated vector and matrix operations which are recognized as intrinsic functions by our AAD software. We compare this approach with naive and vectorized implementations for CPUs. We use four increasingly complex cost functions to evaluate the performance with respect to memory consumption and gradient computation times. Using vectorization, CPU and GPU memory consumption could be substantially reduced compared to the naive reference implementation, in some cases even by an order of complexity. The vectorization allowed usage of optimized parallel libraries during forward and reverse passes which resulted in high speedups for the vectorized CPU version compared to the naive reference implementation. The GPU version achieved an additional speedup of 7.5 ± 4.4, showing that the processing power of GPUs can be utilized for AAD using this concept. Furthermore, we show how this software can be systematically extended for more complex problems such as nonlinear absorption reconstruction for fluorescence-mediated tomography
Adjoint Algorithm for CAD-Based Shape Optimization Using a Cartesian Method
NASA Technical Reports Server (NTRS)
Nemec, Marian; Aftosmis, Michael J.
2004-01-01
Adjoint solutions of the governing flow equations are becoming increasingly important for the development of efficient analysis and optimization algorithms. A well-known use of the adjoint method is gradient-based shape optimization. Given an objective function that defines some measure of performance, such as the lift and drag functionals, its gradient is computed at a cost that is essentially independent of the number of design variables (geometric parameters that control the shape). More recently, emerging adjoint applications focus on the analysis problem, where the adjoint solution is used to drive mesh adaptation, as well as to provide estimates of functional error bounds and corrections. The attractive feature of this approach is that the mesh-adaptation procedure targets a specific functional, thereby localizing the mesh refinement and reducing computational cost. Our focus is on the development of adjoint-based optimization techniques for a Cartesian method with embedded boundaries.12 In contrast t o implementations on structured and unstructured grids, Cartesian methods decouple the surface discretization from the volume mesh. This feature makes Cartesian methods well suited for the automated analysis of complex geometry problems, and consequently a promising approach to aerodynamic optimization. Melvin et developed an adjoint formulation for the TRANAIR code, which is based on the full-potential equation with viscous corrections. More recently, Dadone and Grossman presented an adjoint formulation for the Euler equations. In both approaches, a boundary condition is introduced to approximate the effects of the evolving surface shape that results in accurate gradient computation. Central to automated shape optimization algorithms is the issue of geometry modeling and control. The need to optimize complex, "real-life" geometry provides a strong incentive for the use of parametric-CAD systems within the optimization procedure. In previous work, we presented
An Exact Dual Adjoint Solution Method for Turbulent Flows on Unstructured Grids
NASA Technical Reports Server (NTRS)
Nielsen, Eric J.; Lu, James; Park, Michael A.; Darmofal, David L.
2003-01-01
An algorithm for solving the discrete adjoint system based on an unstructured-grid discretization of the Navier-Stokes equations is presented. The method is constructed such that an adjoint solution exactly dual to a direct differentiation approach is recovered at each time step, yielding a convergence rate which is asymptotically equivalent to that of the primal system. The new approach is implemented within a three-dimensional unstructured-grid framework and results are presented for inviscid, laminar, and turbulent flows. Improvements to the baseline solution algorithm, such as line-implicit relaxation and a tight coupling of the turbulence model, are also presented. By storing nearest-neighbor terms in the residual computation, the dual scheme is computationally efficient, while requiring twice the memory of the flow solution. The scheme is expected to have a broad impact on computational problems related to design optimization as well as error estimation and grid adaptation efforts.
Adjoint-Based Algorithms for Adaptation and Design Optimizations on Unstructured Grids
NASA Technical Reports Server (NTRS)
Nielsen, Eric J.
2006-01-01
Schemes based on discrete adjoint algorithms present several exciting opportunities for significantly advancing the current state of the art in computational fluid dynamics. Such methods provide an extremely efficient means for obtaining discretely consistent sensitivity information for hundreds of design variables, opening the door to rigorous, automated design optimization of complex aerospace configuration using the Navier-Stokes equation. Moreover, the discrete adjoint formulation provides a mathematically rigorous foundation for mesh adaptation and systematic reduction of spatial discretization error. Error estimates are also an inherent by-product of an adjoint-based approach, valuable information that is virtually non-existent in today's large-scale CFD simulations. An overview of the adjoint-based algorithm work at NASA Langley Research Center is presented, with examples demonstrating the potential impact on complex computational problems related to design optimization as well as mesh adaptation.
Solution of the self-adjoint radiative transfer equation on hybrid computer systems
NASA Astrophysics Data System (ADS)
Gasilov, V. A.; Kuchugov, P. A.; Olkhovskaya, O. G.; Chetverushkin, B. N.
2016-06-01
A new technique for simulating three-dimensional radiative energy transfer for the use in the software designed for the predictive simulation of plasma with high energy density on parallel computers is proposed. A highly scalable algorithm that takes into account the angular dependence of the radiation intensity and is free of the ray effect is developed based on the solution of a second-order equation with a self-adjoint operator. A distinctive feature of this algorithm is a preliminary transformation of rotation to eliminate mixed derivatives with respect to the spatial variables, simplify the structure of the difference operator, and accelerate the convergence of the iterative solution of the equation. It is shown that the proposed method correctly reproduces the limiting cases—isotropic radiation and the directed radiation with a δ-shaped angular distribution.
Heberton, C.I.; Russell, T.F.; Konikow, L.F.; Hornberger, G.Z.
2000-01-01
This report documents the U.S. Geological Survey Eulerian-Lagrangian Localized Adjoint Method (ELLAM) algorithm that solves an integral form of the solute-transport equation, incorporating an implicit-in-time difference approximation for the dispersive and sink terms. Like the algorithm in the original version of the U.S. Geological Survey MOC3D transport model, ELLAM uses a method of characteristics approach to solve the transport equation on the basis of the velocity field. The ELLAM algorithm, however, is based on an integral formulation of conservation of mass and uses appropriate numerical techniques to obtain global conservation of mass. The implicit procedure eliminates several stability criteria required for an explicit formulation. Consequently, ELLAM allows large transport time increments to be used. ELLAM can produce qualitatively good results using a small number of transport time steps. A description of the ELLAM numerical method, the data-input requirements and output options, and the results of simulator testing and evaluation are presented. The ELLAM algorithm was evaluated for the same set of problems used to test and evaluate Version 1 and Version 2 of MOC3D. These test results indicate that ELLAM offers a viable alternative to the explicit and implicit solvers in MOC3D. Its use is desirable when mass balance is imperative or a fast, qualitative model result is needed. Although accurate solutions can be generated using ELLAM, its efficiency relative to the two previously documented solution algorithms is problem dependent.
Nonlinear self adjointness, conservation laws and exact solutions of ill-posed Boussinesq equation
NASA Astrophysics Data System (ADS)
Yaşar, Emrullah; San, Sait; Özkan, Yeşim Sağlam
2016-01-01
In this work, we consider the ill-posed Boussinesq equation which arises in shallow water waves and non-linear lattices. We prove that the ill-posed Boussinesq equation is nonlinearly self-adjoint. Using this property and Lie point symmetries, we construct conservation laws for the underlying equation. In addition, the generalized solitonary, periodic and compact-like solutions are constructed by the exp-function method.
NASA Technical Reports Server (NTRS)
Pulliam, T. H.; Nemec, M.; Holst, T.; Zingg, D. W.; Kwak, Dochan (Technical Monitor)
2002-01-01
A comparison between an Evolutionary Algorithm (EA) and an Adjoint-Gradient (AG) Method applied to a two-dimensional Navier-Stokes code for airfoil design is presented. Both approaches use a common function evaluation code, the steady-state explicit part of the code,ARC2D. The parameterization of the design space is a common B-spline approach for an airfoil surface, which together with a common griding approach, restricts the AG and EA to the same design space. Results are presented for a class of viscous transonic airfoils in which the optimization tradeoff between drag minimization as one objective and lift maximization as another, produces the multi-objective design space. Comparisons are made for efficiency, accuracy and design consistency.
NASA Astrophysics Data System (ADS)
Al-Attar, D.; Woodhouse, J. H.
2011-12-01
Normal mode spectra provide a valuable data set for global seismic tomography, and, notably, are amongst the few geophysical observables that are sensitive to lateral variations in density structure within the Earth. Nonetheless, the effects of lateral density variations on mode spectra are rather subtle. In order, therefore, to reliably determine density variations with in the earth, it is necessary to make use of sufficiently accurate methods for calculating synthetic mode spectra. In particular, recent work has highlighted the need to perform 'full-coupling calculations' that take into account the interaction of large numbers of spherical earth multiplets. However, present methods for performing such full-coupling calculations require diagonalization of large coupling matrices, and so become computationally inefficient as the number of coupled modes is increased. In order to perform full-coupling calculations more efficiently, we describe a new implementation of the direct solution method for calculating synthetic spectra in laterally heterogeneous earth models. This approach is based on the solution of the inhomogeneous mode coupling equations in the frequency domain, and does not require the diagonalization of large matrices. Early implementations of the direct solution method used LU-decomposition to solve the mode coupling equations. However, as the number of coupled modes is increased, this method becomes impractically slow. To circumvent this problem, we solve the mode coupling equations iteratively using the preconditioned biconjugate gradient algorithm. We present a number of numerical tests to display the accuracy and efficiency of this method for performing large full-coupling calculations. In addition, we describe a frequency-domain formulation of the adjoint method for the calculation of Frechet kernels that show the sensitivity of normal mode observations to variations in earth structure. The calculation of such Frechet kernels involves one solution
Spectral Solutions of Self-adjoint Elliptic Problems with Immersed Interfaces
Auchmuty, G.; Kloucek, P.
2011-12-15
This paper describes a spectral representation of solutions of self-adjoint elliptic problems with immersed interfaces. The interface is assumed to be a simple non-self-intersecting closed curve that obeys some weak regularity conditions. The problem is decomposed into two problems, one with zero interface data and the other with zero exterior boundary data. The problem with zero interface data is solved by standard spectral methods. The problem with non-zero interface data is solved by introducing an interface space H{sub {Gamma}}({Omega}) and constructing an orthonormal basis of this space. This basis is constructed using a special class of orthogonal eigenfunctions analogously to the methods used for standard trace spaces by Auchmuty (SIAM J. Math. Anal. 38, 894-915, 2006). Analytical and numerical approximations of these eigenfunctions are described and some simulations are presented.
Haydock’s recursive solution of self-adjoint problems. Discrete spectrum
Moroz, Alexander
2014-12-15
Haydock’s recursive solution is shown to underline a number of different concepts such as (i) quasi-exactly solvable models, (ii) exactly solvable models, (iii) three-term recurrence solutions based on Schweber’s quantization criterion in Hilbert spaces of entire analytic functions, and (iv) a discrete quantum mechanics of Odake and Sasaki. A recurrent theme of Haydock’s recursive solution is that the spectral properties of any self-adjoint problem can be mapped onto a corresponding sequence of polynomials (p{sub n}(E)) in energy variable E. The polynomials (p{sub n}(E)) are orthonormal with respect to the density of states n{sub 0}(E) and energy eigenstate |E〉 is the generating function of (p{sub n}(E)). The generality of Haydock’s recursive solution enables one to see the different concepts from a unified perspective and mutually benefiting from each other. Some results obtained within the particular framework of any of (i) to (iv) may have much broader significance.
Southern California Adjoint Source Inversions
NASA Astrophysics Data System (ADS)
Tromp, J.; Kim, Y.
2007-12-01
Southern California Centroid-Moment Tensor (CMT) solutions with 9 components (6 moment tensor elements, latitude, longitude, and depth) are sought to minimize a misfit function computed from waveform differences. The gradient of a misfit function is obtained based upon two numerical simulations for each earthquake: one forward calculation for the southern California model, and an adjoint calculation that uses time-reversed signals at the receivers. Conjugate gradient and square-root variable metric methods are used to iteratively improve the earthquake source model while reducing the misfit function. The square-root variable metric algorithm has the advantage of providing a direct approximation to the posterior covariance operator. We test the inversion procedure by perturbing each component of the CMT solution, and see how the algorithm converges. Finally, we demonstrate full inversion capabilities using data for real Southern California earthquakes.
Healy, R.W.; Russell, T.F.
1993-01-01
Test results demonstrate that the finite-volume Eulerian-Lagrangian localized adjoint method (FVELLAM) outperforms standard finite-difference methods for solute transport problems that are dominated by advection. FVELLAM systematically conserves mass globally with all types of boundary conditions. Integrated finite differences, instead of finite elements, are used to approximate the governing equation. This approach, in conjunction with a forward tracking scheme, greatly facilitates mass conservation. The mass storage integral is numerically evaluated at the current time level, and quadrature points are then tracked forward in time to the next level. Forward tracking permits straightforward treatment of inflow boundaries, thus avoiding the inherent problem in backtracking of characteristic lines intersecting inflow boundaries. FVELLAM extends previous results by obtaining mass conservation locally on Lagrangian space-time elements. -from Authors
Healy, R.W.; Russell, T.F.
1992-01-01
A finite-volume Eulerian-Lagrangian local adjoint method for solution of the advection-dispersion equation is developed and discussed. The method is mass conservative and can solve advection-dominated ground-water solute-transport problems accurately and efficiently. An integrated finite-difference approach is used in the method. A key component of the method is that the integral representing the mass-storage term is evaluated numerically at the current time level. Integration points, and the mass associated with these points, are then forward tracked up to the next time level. The number of integration points required to reach a specified level of accuracy is problem dependent and increases as the sharpness of the simulated solute front increases. Integration points are generally equally spaced within each grid cell. For problems involving variable coefficients it has been found to be advantageous to include additional integration points at strategic locations in each well. These locations are determined by backtracking. Forward tracking of boundary fluxes by the method alleviates problems that are encountered in the backtracking approaches of most characteristic methods. A test problem is used to illustrate that the new method offers substantial advantages over other numerical methods for a wide range of problems.
Mesh-free adjoint methods for nonlinear filters
NASA Astrophysics Data System (ADS)
Daum, Fred
2005-09-01
We apply a new industrial strength numerical approximation, called the "mesh-free adjoint method", to solve the nonlinear filtering problem. This algorithm exploits the smoothness of the problem, unlike particle filters, and hence we expect that mesh-free adjoints are superior to particle filters for many practical applications. The nonlinear filter problem is equivalent to solving the Fokker-Planck equation in real time. The key idea is to use a good adaptive non-uniform quantization of state space to approximate the solution of the Fokker-Planck equation. In particular, the adjoint method computes the location of the nodes in state space to minimize errors in the final answer. This use of an adjoint is analogous to optimal control algorithms, but it is more interesting. The adjoint method is also analogous to importance sampling in particle filters, but it is better for four reasons: (1) it exploits the smoothness of the problem; (2) it explicitly minimizes the errors in the relevant functional; (3) it explicitly models the dynamics in state space; and (4) it can be used to compute a corrected value for the desired functional using the residuals. We will attempt to make this paper accessible to normal engineers who do not have PDEs for breakfast.
Adjoint operator approach to shape design for internal incompressible flows
NASA Technical Reports Server (NTRS)
Cabuk, H.; Sung, C.-H.; Modi, V.
1991-01-01
The problem of determining the profile of a channel or duct that provides the maximum static pressure rise is solved. Incompressible, laminar flow governed by the steady state Navier-Stokes equations is assumed. Recent advances in computational resources and algorithms have made it possible to solve the direct problem of determining such a flow through a body of known geometry. It is possible to obtain a set of adjoint equations, the solution to which permits the calculation of the direction and relative magnitude of change in the diffuser profile that leads to a higher pressure rise. The solution to the adjoint problem can be shown to represent an artificially constructed flow. This interpretation provides a means to construct numerical solutions to the adjoint equations that do not compromise the fully viscous nature of the problem. The algorithmic and computational aspects of solving the adjoint equations are addressed. The form of these set of equations is similar but not identical to the Navier-Stokes equations. In particular some issues related to boundary conditions and stability are discussed.
Adjoint Error Estimation for Linear Advection
Connors, J M; Banks, J W; Hittinger, J A; Woodward, C S
2011-03-30
An a posteriori error formula is described when a statistical measurement of the solution to a hyperbolic conservation law in 1D is estimated by finite volume approximations. This is accomplished using adjoint error estimation. In contrast to previously studied methods, the adjoint problem is divorced from the finite volume method used to approximate the forward solution variables. An exact error formula and computable error estimate are derived based on an abstractly defined approximation of the adjoint solution. This framework allows the error to be computed to an arbitrary accuracy given a sufficiently well resolved approximation of the adjoint solution. The accuracy of the computable error estimate provably satisfies an a priori error bound for sufficiently smooth solutions of the forward and adjoint problems. The theory does not currently account for discontinuities. Computational examples are provided that show support of the theory for smooth solutions. The application to problems with discontinuities is also investigated computationally.
Introduction to Adjoint Models
NASA Technical Reports Server (NTRS)
Errico, Ronald M.
2015-01-01
In this lecture, some fundamentals of adjoint models will be described. This includes a basic derivation of tangent linear and corresponding adjoint models from a parent nonlinear model, the interpretation of adjoint-derived sensitivity fields, a description of methods of automatic differentiation, and the use of adjoint models to solve various optimization problems, including singular vectors. Concluding remarks will attempt to correct common misconceptions about adjoint models and their utilization.
Hekmat, Mohamad Hamed; Mirzaei, Masoud
2015-01-01
In the present research, we tried to improve the performance of the lattice Boltzmann (LB) -based adjoint approach by utilizing the mesoscopic inherent of the LB method. In this regard, two macroscopic discrete adjoint (MADA) and microscopic discrete adjoint (MIDA) approaches are used to answer the following two challenging questions. Is it possible to extend the concept of the macroscopic and microscopic variables of the flow field to the corresponding adjoint ones? Further, similar to the conservative laws in the LB method, is it possible to find the comparable conservation equations in the adjoint approach? If so, then a definite framework, similar to that used in the flow solution by the LB method, can be employed in the flow sensitivity analysis by the MIDA approach. This achievement can decrease the implementation cost and coding efforts of the MIDA method in complicated sensitivity analysis problems. First, the MADA and MIDA equations are extracted based on the LB method using the duality viewpoint. Meanwhile, using an elementary case, inverse design of a two-dimensional unsteady Poiseuille flow in a periodic channel with constant body forces, the procedure of analytical evaluation of the adjoint variables is described. The numerical results show that similar correlations between the distribution functions can be seen between the corresponding adjoint ones. Besides, the results are promising, emphasizing the flow field adjoint variables can be evaluated via the adjoint distribution functions. Finally, the adjoint conservative laws are introduced. PMID:25679735
Adjoint Data Assimilative Model Study of the Gulf of Maine Coastal Circulation
NASA Astrophysics Data System (ADS)
He, R.; McGillicuddy, D. J.; Lynch, D. R.
2004-12-01
Data assimilation (DA) in the coastal ocean can be divided into category of either sequential estimation or variational adjoint. Sequential estimation techniques blend models with observations directly, using a variety of algorithms with which the relative weights of data and model are calculated. Variational adjoint techniques infer model control variables (e.g. parameters, forcing functions, boundary conditions, etc.) that minimize the misfit between observations and predictions. The advantage of the latter techniques over the former is that the resulting model solutions obey model dynamics. In this study, the Gulf of Maine coastal circulation and the material property transport are investigated with the Dartmouth variational adjoint DA modeling system, which assimilates in-situ data via inversion for the unknown sea level elevation at open boundaries. In-situ observations include ADCP currents and coastal sea levels. The adjoint DA model skill is evaluated by the inter-comparisons between modeled and observed drifter trajectories. Excellent model skill is found, demonstrating the utility and effectiveness of the adjoint DA modeling system in bridging in-situ observations with coastal ocean model simulations. Implications of the adjoint DA strategy on the emergent coastal ocean observing systems are discussed.
Healy, R.W.; Russell, T.F.
1998-01-01
We extend the finite-volume Eulerian-Lagrangian localized adjoint method (FVELLAM) for solution of the advection-dispersion equation to two dimensions. The method can conserve mass globally and is not limited by restrictions on the size of the grid Peclet or Courant number. Therefore, it is well suited for solution of advection-dominated ground-water solute transport problems. In test problem comparisons with standard finite differences, FVELLAM is able to attain accurate solutions on much coarser space and time grids. On fine grids, the accuracy of the two methods is comparable. A critical aspect of FVELLAM (and all other ELLAMs) is evaluation of the mass storage integral from the preceding time level. In FVELLAM this may be accomplished with either a forward or backtracking approach. The forward tracking approach conserves mass globally and is the preferred approach. The backtracking approach is less computationally intensive, but not globally mass conservative. Boundary terms are systematically represented as integrals in space and time which are evaluated by a common integration scheme in conjunction with forward tracking through time. Unlike the one-dimensional case, local mass conservation cannot be guaranteed, so slight oscillations in concentration can develop, particularly in the vicinity of inflow or outflow boundaries. Published by Elsevier Science Ltd.
NASA Astrophysics Data System (ADS)
Bozdag, Ebru; Lefebvre, Matthieu; Lei, Wenjie; Peter, Daniel; Smith, James; Komatitsch, Dimitri; Tromp, Jeroen
2015-04-01
We will present our initial results of global adjoint tomography based on 3D seismic wave simulations which is one of the most challenging examples in seismology in terms of intense computational requirements and vast amount of high-quality seismic data that can potentially be assimilated in inversions. Using a spectral-element method, we incorporate full 3D wave propagation in seismic tomography by running synthetic seismograms and adjoint simulations to compute exact sensitivity kernels in realistic 3D background models. We run our global simulations on the Oak Ridge National Laboratory's Cray XK7 "Titan" system taking advantage of the GPU version of the SPECFEM3D_GLOBE package. We have started iterations with initially selected 253 earthquakes within the magnitude range of 5.5 < Mw < 7.0 and numerical simulations having resolution down to ~27 s to invert for a transversely isotropic crust and mantle model using a non-linear conjugate gradient algorithm. The measurements are currently based on frequency-dependent traveltime misfits. We use both minor- and major-arc body and surface waves by running 200 min simulations where inversions are performed with more than 2.6 million measurements. Our initial results after 12 iterations already indicate several prominent features such as enhanced slab (e.g., Hellenic, Japan, Bismarck, Sandwich), plume/hotspot (e.g., the Pacific superplume, Caroline, Yellowstone, Hawaii) images, etc. To improve the resolution and ray coverage, particularly in the lower mantle, our aim is to increase the resolution of numerical simulations first going down to ~17 s and then to ~9 s to incorporate high-frequency body waves in inversions. While keeping track of the progress and illumination of features in our models with a limited data set, we work towards to assimilate all available data in inversions from all seismic networks and earthquakes in the global CMT catalogue.
Using an Adjoint Approach to Eliminate Mesh Sensitivities in Computational Design
NASA Technical Reports Server (NTRS)
Nielsen, Eric J.; Park, Michael A.
2006-01-01
An algorithm for efficiently incorporating the effects of mesh sensitivities in a computational design framework is introduced. The method is based on an adjoint approach and eliminates the need for explicit linearizations of the mesh movement scheme with respect to the geometric parameterization variables, an expense that has hindered practical large-scale design optimization using discrete adjoint methods. The effects of the mesh sensitivities can be accounted for through the solution of an adjoint problem equivalent in cost to a single mesh movement computation, followed by an explicit matrix-vector product scaling with the number of design variables and the resolution of the parameterized surface grid. The accuracy of the implementation is established and dramatic computational savings obtained using the new approach are demonstrated using several test cases. Sample design optimizations are also shown.
Using an Adjoint Approach to Eliminate Mesh Sensitivities in Computational Design
NASA Technical Reports Server (NTRS)
Nielsen, Eric J.; Park, Michael A.
2005-01-01
An algorithm for efficiently incorporating the effects of mesh sensitivities in a computational design framework is introduced. The method is based on an adjoint approach and eliminates the need for explicit linearizations of the mesh movement scheme with respect to the geometric parameterization variables, an expense that has hindered practical large-scale design optimization using discrete adjoint methods. The effects of the mesh sensitivities can be accounted for through the solution of an adjoint problem equivalent in cost to a single mesh movement computation, followed by an explicit matrix-vector product scaling with the number of design variables and the resolution of the parameterized surface grid. The accuracy of the implementation is established and dramatic computational savings obtained using the new approach are demonstrated using several test cases. Sample design optimizations are also shown.
Adjoint Formulation for an Embedded-Boundary Cartesian Method
NASA Technical Reports Server (NTRS)
Nemec, Marian; Aftosmis, Michael J.; Murman, Scott M.; Pulliam, Thomas H.
2004-01-01
Many problems in aerodynamic design can be characterized by smooth and convex objective functions. This motivates the use of gradient-based algorithms, particularly for problems with a large number of design variables, to efficiently determine optimal shapes and configurations that maximize aerodynamic performance. Accurate and efficient computation of the gradient, however, remains a challenging task. In optimization problems where the number of design variables dominates the number of objectives and flow- dependent constraints, the cost of gradient computations can be significantly reduced by the use of the adjoint method. The problem of aerodynamic optimization using the adjoint method has been analyzed and validated for both structured and unstructured grids. The method has been applied to design problems governed by the potential, Euler, and Navier-Stokes equations and can be subdivided into the continuous and discrete formulations. Giles and Pierce provide a detailed review of both approaches. Most implementations rely on grid-perturbation or mapping procedures during the gradient computation that explicitly couple changes in the surface shape to the volume grid. The solution of the adjoint equation is usually accomplished using the same scheme that solves the governing flow equations. Examples of such code reuse include multistage Runge-Kutta schemes coupled with multigrid, approximate-factorization, line-implicit Gauss-Seidel, and also preconditioned GMRES. The development of the adjoint method for aerodynamic optimization problems on Cartesian grids has been limited. In contrast to implementations on structured and unstructured grids, Cartesian grid methods decouple the surface discretization from the volume grid. This feature makes Cartesian methods well suited for the automated analysis of complex geometry problems, and consequently a promising approach to aerodynamic optimization. Melvin e t al. developed an adjoint formulation for the TRANAIR code
Generalized uncertainty principle and self-adjoint operators
Balasubramanian, Venkat; Das, Saurya; Vagenas, Elias C.
2015-09-15
In this work we explore the self-adjointness of the GUP-modified momentum and Hamiltonian operators over different domains. In particular, we utilize the theorem by von-Neumann for symmetric operators in order to determine whether the momentum and Hamiltonian operators are self-adjoint or not, or they have self-adjoint extensions over the given domain. In addition, a simple example of the Hamiltonian operator describing a particle in a box is given. The solutions of the boundary conditions that describe the self-adjoint extensions of the specific Hamiltonian operator are obtained.
Towards Global Adjoint Tomography
NASA Astrophysics Data System (ADS)
Bozdag, E.; Zhu, H.; Peter, D. B.; Tromp, J.
2012-12-01
Seismic tomography is at a stage where we can harness entire seismograms using the opportunities offered by advances in numerical wave propagation solvers and high-performance computing. Adjoint methods provide an efficient way for incorporating full nonlinearity of wave propagation and 3D Fréchet kernels in iterative seismic inversions which have so far given promising results at continental and regional scales. Our goal is to take adjoint tomography forward to image the entire planet. Using an iterative conjugate gradient scheme, we initially set the aim to obtain a global crustal and mantle model with confined transverse isotropy in the upper mantle. We have started with around 255 global CMT events having moment magnitudes between 5.8 and 7, and used GSN stations as well as some local networks such as USArray, European stations etc. Prior to the structure inversion, we reinvert global CMT solutions by computing Green functions in our 3D reference model to take into account effects of crustal variations on source parameters. Using the advantages of numerical simulations, our strategy is to invert crustal and mantle structure together to avoid any bias introduced into upper-mantle images due to "crustal corrections", which are commonly used in classical tomography. 3D simulations dramatically increase the usable amount of data so that, with the current earthquake-station setup, we perform each iteration with more than two million measurements. Multi-resolution smoothing based on ray density is applied to the gradient to better deal with the imperfect source-station distribution on the globe and extract more information underneath regions with dense ray coverage and vice versa. Similar to frequency domain approach, we reduce nonlinearities by starting from long periods and gradually increase the frequency content of data after successive model updates. To simplify the problem, we primarily focus on the elastic structure and therefore our measurements are based on
Towards efficient backward-in-time adjoint computations using data compression techniques
Cyr, E. C.; Shadid, J. N.; Wildey, T.
2014-12-16
In the context of a posteriori error estimation for nonlinear time-dependent partial differential equations, the state-of-the-practice is to use adjoint approaches which require the solution of a backward-in-time problem defined by a linearization of the forward problem. One of the major obstacles in the practical application of these approaches, we found, is the need to store, or recompute, the forward solution to define the adjoint problem and to evaluate the error representation. Our study considers the use of data compression techniques to approximate forward solutions employed in the backward-in-time integration. The development derives an error representation that accounts for themore » difference between the standard-approach and the compressed approximation of the forward solution. This representation is algorithmically similar to the standard representation and only requires the computation of the quantity of interest for the forward solution and the data-compressed reconstructed solution (i.e. scalar quantities that can be evaluated as the forward problem is integrated). This approach is then compared with existing techniques, such as checkpointing and time-averaged adjoints. Lastly, we provide numerical results indicating the potential efficiency of our approach on a transient diffusion–reaction equation and on the Navier–Stokes equations. These results demonstrate memory compression ratios up to 450×450× while maintaining reasonable accuracy in the error-estimates.« less
Towards efficient backward-in-time adjoint computations using data compression techniques
Cyr, E. C.; Shadid, J. N.; Wildey, T.
2014-12-16
In the context of a posteriori error estimation for nonlinear time-dependent partial differential equations, the state-of-the-practice is to use adjoint approaches which require the solution of a backward-in-time problem defined by a linearization of the forward problem. One of the major obstacles in the practical application of these approaches, we found, is the need to store, or recompute, the forward solution to define the adjoint problem and to evaluate the error representation. Our study considers the use of data compression techniques to approximate forward solutions employed in the backward-in-time integration. The development derives an error representation that accounts for the difference between the standard-approach and the compressed approximation of the forward solution. This representation is algorithmically similar to the standard representation and only requires the computation of the quantity of interest for the forward solution and the data-compressed reconstructed solution (i.e. scalar quantities that can be evaluated as the forward problem is integrated). This approach is then compared with existing techniques, such as checkpointing and time-averaged adjoints. Lastly, we provide numerical results indicating the potential efficiency of our approach on a transient diffusion–reaction equation and on the Navier–Stokes equations. These results demonstrate memory compression ratios up to 450×450× while maintaining reasonable accuracy in the error-estimates.
Fully automatic adjoints: a robust and efficient mechanism for generating adjoint ocean models
NASA Astrophysics Data System (ADS)
Ham, D. A.; Farrell, P. E.; Funke, S. W.; Rognes, M. E.
2012-04-01
The problem of generating and maintaining adjoint models is sufficiently difficult that typically only the most advanced and well-resourced community ocean models achieve it. There are two current technologies which each suffer from their own limitations. Algorithmic differentiation, also called automatic differentiation, is employed by models such as the MITGCM [2] and the Alfred Wegener Institute model FESOM [3]. This technique is very difficult to apply to existing code, and requires a major initial investment to prepare the code for automatic adjoint generation. AD tools may also have difficulty with code employing modern software constructs such as derived data types. An alternative is to formulate the adjoint differential equation and to discretise this separately. This approach, known as the continuous adjoint and employed in ROMS [4], has the disadvantage that two different model code bases must be maintained and manually kept synchronised as the model develops. The discretisation of the continuous adjoint is not automatically consistent with that of the forward model, producing an additional source of error. The alternative presented here is to formulate the flow model in the high level language UFL (Unified Form Language) and to automatically generate the model using the software of the FEniCS project. In this approach it is the high level code specification which is differentiated, a task very similar to the formulation of the continuous adjoint [5]. However since the forward and adjoint models are generated automatically, the difficulty of maintaining them vanishes and the software engineering process is therefore robust. The scheduling and execution of the adjoint model, including the application of an appropriate checkpointing strategy is managed by libadjoint [1]. In contrast to the conventional algorithmic differentiation description of a model as a series of primitive mathematical operations, libadjoint employs a new abstraction of the simulation
An ordinary differential equation based solution path algorithm.
Wu, Yichao
2011-01-01
Efron, Hastie, Johnstone and Tibshirani (2004) proposed Least Angle Regression (LAR), a solution path algorithm for the least squares regression. They pointed out that a slight modification of the LAR gives the LASSO (Tibshirani, 1996) solution path. However it is largely unknown how to extend this solution path algorithm to models beyond the least squares regression. In this work, we propose an extension of the LAR for generalized linear models and the quasi-likelihood model by showing that the corresponding solution path is piecewise given by solutions of ordinary differential equation systems. Our contribution is twofold. First, we provide a theoretical understanding on how the corresponding solution path propagates. Second, we propose an ordinary differential equation based algorithm to obtain the whole solution path. PMID:21532936
Application of adjoint operators to neural learning
NASA Technical Reports Server (NTRS)
Barhen, J.; Toomarian, N.; Gulati, S.
1990-01-01
A technique for the efficient analytical computation of such parameters of the neural architecture as synaptic weights and neural gain is presented as a single solution of a set of adjoint equations. The learning model discussed concentrates on the adiabatic approximation only. A problem of interest is represented by a system of N coupled equations, and then adjoint operators are introduced. A neural network is formalized as an adaptive dynamical system whose temporal evolution is governed by a set of coupled nonlinear differential equations. An approach based on the minimization of a constrained neuromorphic energylike function is applied, and the complete learning dynamics are obtained as a result of the calculations.
Diagnositcs With Adjoint Modelling
NASA Astrophysics Data System (ADS)
Blessing, S.; Fraedrich, K.; Kirk, E.; Lunkeit, F.
The potential usefulness of an adjoint primitive equations global atmospheric circu- lation model for climate diagnostics is demonstrated in a feasibility study. A daily NAO-type index is calculated as one-point correlation of the 300 hPa streamfunction anomaly. By application of the adjoint model we diagnose its temperature forcing on short timescales in terms of spatial temperature sensitivity patterns at different time lags, which, in a first order approximation, induce growth of the index. The dynamical relevance of these sensitivity patterns is confirmed by lag-correlating the index time series and the projection time series of the model temperature on these sensitivity patterns.
Adjoint affine fusion and tadpoles
NASA Astrophysics Data System (ADS)
Urichuk, Andrew; Walton, Mark A.
2016-06-01
We study affine fusion with the adjoint representation. For simple Lie algebras, elementary and universal formulas determine the decomposition of a tensor product of an integrable highest-weight representation with the adjoint representation. Using the (refined) affine depth rule, we prove that equally striking results apply to adjoint affine fusion. For diagonal fusion, a coefficient equals the number of nonzero Dynkin labels of the relevant affine highest weight, minus 1. A nice lattice-polytope interpretation follows and allows the straightforward calculation of the genus-1 1-point adjoint Verlinde dimension, the adjoint affine fusion tadpole. Explicit formulas, (piecewise) polynomial in the level, are written for the adjoint tadpoles of all classical Lie algebras. We show that off-diagonal adjoint affine fusion is obtained from the corresponding tensor product by simply dropping non-dominant representations.
Ayyoubzadeh, Seyed Mohsen; Vosoughi, Naser
2011-09-14
Obtaining the set of algebraic equations that directly correspond to a physical phenomenon has been viable in the recent direct discrete method (DDM). Although this method may find its roots in physical and geometrical considerations, there are still some degrees of freedom that one may suspect optimize-able. Here we have used the information embedded in the corresponding adjoint equation to form a local functional, which in turn by its minimization, yield suitable dual mesh positioning.
Adjoint-based uncertainty quantification and sensitivity analysis for reactor depletion calculations
NASA Astrophysics Data System (ADS)
Stripling, Hayes Franklin
Depletion calculations for nuclear reactors model the dynamic coupling between the material composition and neutron flux and help predict reactor performance and safety characteristics. In order to be trusted as reliable predictive tools and inputs to licensing and operational decisions, the simulations must include an accurate and holistic quantification of errors and uncertainties in its outputs. Uncertainty quantification is a formidable challenge in large, realistic reactor models because of the large number of unknowns and myriad sources of uncertainty and error. We present a framework for performing efficient uncertainty quantification in depletion problems using an adjoint approach, with emphasis on high-fidelity calculations using advanced massively parallel computing architectures. This approach calls for a solution to two systems of equations: (a) the forward, engineering system that models the reactor, and (b) the adjoint system, which is mathematically related to but different from the forward system. We use the solutions of these systems to produce sensitivity and error estimates at a cost that does not grow rapidly with the number of uncertain inputs. We present the framework in a general fashion and apply it to both the source-driven and k-eigenvalue forms of the depletion equations. We describe the implementation and verification of solvers for the forward and ad- joint equations in the PDT code, and we test the algorithms on realistic reactor analysis problems. We demonstrate a new approach for reducing the memory and I/O demands on the host machine, which can be overwhelming for typical adjoint algorithms. Our conclusion is that adjoint depletion calculations using full transport solutions are not only computationally tractable, they are the most attractive option for performing uncertainty quantification on high-fidelity reactor analysis problems.
Organization mechanism and counting algorithm on vertex-cover solutions
NASA Astrophysics Data System (ADS)
Wei, Wei; Zhang, Renquan; Niu, Baolong; Guo, Binghui; Zheng, Zhiming
2015-04-01
Counting the solution number of combinational optimization problems is an important topic in the study of computational complexity, which is concerned with Vertex-Cover in this paper. First, we investigate organizations of Vertex-Cover solution spaces by the underlying connectivity of unfrozen vertices and provide facts on the global and local environment. Then, a Vertex-Cover Solution Number Counting Algorithm is proposed and its complexity analysis is provided, the results of which fit very well with the simulations and have a better performance than those by 1-RSB in the neighborhood of c = e for random graphs. Based on the algorithm, variation and fluctuation on the solution number the statistics are studied to reveal the evolution mechanism of the solution numbers. Furthermore, the marginal probability distributions on the solution space are investigated on both the random graph and scale-free graph to illustrate the different evolution characteristics of their solution spaces. Thus, doing solution number counting based on the graph expression of the solution space should be an alternative and meaningful way to study the hardness of NP-complete and #P-complete problems and the appropriate algorithm design can help to achieve better approximations of solving combinational optimization problems and the corresponding counting problems.
A solution quality assessment method for swarm intelligence optimization algorithms.
Zhang, Zhaojun; Wang, Gai-Ge; Zou, Kuansheng; Zhang, Jianhua
2014-01-01
Nowadays, swarm intelligence optimization has become an important optimization tool and wildly used in many fields of application. In contrast to many successful applications, the theoretical foundation is rather weak. Therefore, there are still many problems to be solved. One problem is how to quantify the performance of algorithm in finite time, that is, how to evaluate the solution quality got by algorithm for practical problems. It greatly limits the application in practical problems. A solution quality assessment method for intelligent optimization is proposed in this paper. It is an experimental analysis method based on the analysis of search space and characteristic of algorithm itself. Instead of "value performance," the "ordinal performance" is used as evaluation criteria in this method. The feasible solutions were clustered according to distance to divide solution samples into several parts. Then, solution space and "good enough" set can be decomposed based on the clustering results. Last, using relative knowledge of statistics, the evaluation result can be got. To validate the proposed method, some intelligent algorithms such as ant colony optimization (ACO), particle swarm optimization (PSO), and artificial fish swarm algorithm (AFS) were taken to solve traveling salesman problem. Computational results indicate the feasibility of proposed method. PMID:25013845
MCNP: Multigroup/adjoint capabilities
Wagner, J.C.; Redmond, E.L. II; Palmtag, S.P.; Hendricks, J.S.
1994-04-01
This report discusses various aspects related to the use and validity of the general purpose Monte Carlo code MCNP for multigroup/adjoint calculations. The increased desire to perform comparisons between Monte Carlo and deterministic codes, along with the ever-present desire to increase the efficiency of large MCNP calculations has produced a greater user demand for the multigroup/adjoint capabilities. To more fully utilize these capabilities, we review the applications of the Monte Carlo multigroup/adjoint method, describe how to generate multigroup cross sections for MCNP with the auxiliary CRSRD code, describe how to use the multigroup/adjoint capability in MCNP, and provide examples and results indicating the effectiveness and validity of the MCNP multigroup/adjoint treatment. This information should assist users in taking advantage of the MCNP multigroup/adjoint capabilities.
Adjoint-Operator Learning For A Neural Network
NASA Technical Reports Server (NTRS)
Barhen, Jacob; Toomarian, Nikzad
1993-01-01
Electronic neural networks made to synthesize initially unknown mathematical models of time-dependent phenomena or to learn temporally evolving patterns by use of algorithms based on adjoint operators. Algorithms less complicated, involve less computation and solve learning equations forward in time possibly simultaneously with equations of evolution of neural network, thereby both increasing computational efficiency and making real-time applications possible.
A genetic algorithm solution to the unit commitment problem
Kazarlis, S.A.; Bakirtzis, A.G.; Petridis, V.
1996-02-01
This paper presents a Genetic Algorithm (GA) solution to the Unit Commitment problem. GAs are general purpose optimization techniques based on principles inspired from the biological evolution using metaphors of mechanisms such as natural selection, genetic recombination and survival of the fittest. A simple Ga algorithm implementation using the standard crossover and mutation operators could locate near optimal solutions but in most cases failed to converge to the optimal solution. However, using the Varying Quality Function technique and adding problem specific operators, satisfactory solutions to the Unit Commitment problem were obtained. Test results for systems of up to 100 units and comparisons with results obtained using Lagrangian Relaxation and Dynamic Programming are also reported.
Dual of QCD with one adjoint fermion
Mojaza, Matin; Nardecchia, Marco; Pica, Claudio; Sannino, Francesco
2011-03-15
We construct the magnetic dual of QCD with one adjoint Weyl fermion. The dual is a consistent solution of the 't Hooft anomaly matching conditions, allows for flavor decoupling, and remarkably constitutes the first nonsupersymmetric dual valid for any number of colors. The dual allows to bound the anomalous dimension of the Dirac fermion mass operator to be less than one in the conformal window.
Forward and adjoint sensitivity computation of chaotic dynamical systems
Wang, Qiqi
2013-02-15
This paper describes a forward algorithm and an adjoint algorithm for computing sensitivity derivatives in chaotic dynamical systems, such as the Lorenz attractor. The algorithms compute the derivative of long time averaged “statistical” quantities to infinitesimal perturbations of the system parameters. The algorithms are demonstrated on the Lorenz attractor. We show that sensitivity derivatives of statistical quantities can be accurately estimated using a single, short trajectory (over a time interval of 20) on the Lorenz attractor.
Unsteady adjoint of a gas turbine inlet guide vane
NASA Astrophysics Data System (ADS)
Talnikar, Chaitanya; Wang, Qiqi
2015-11-01
Unsteady fluid flow simulations like large eddy simulation have been shown to be crucial in accurately predicting heat transfer in turbomachinery applications like transonic flow over an inlet guide vane. To compute sensitivities of aerothermal objectives for a vane with respect to design parameters an unsteady adjoint is required. In this talk we present unsteady adjoint solutions for a vane from VKI using pressure loss and heat transfer over the vane surface as the objectives. The boundary layer on the suction side near the trailing edge of the vane is turbulent and this poses a challenge for an adjoint solver. The chaotic dynamics cause the adjoint solution to diverge exponentially to infinity from that region when simulated backwards in time. The prospect of adding artificial viscosity to the adjoint equations to dampen the adjoint fields is investigated. Results for the vane from simulations performed on the Titan supercomputer will be shown and the effect of the additional viscosity on the accuracy of the sensitivities will be discussed.
On the Multilevel Solution Algorithm for Markov Chains
NASA Technical Reports Server (NTRS)
Horton, Graham
1997-01-01
We discuss the recently introduced multilevel algorithm for the steady-state solution of Markov chains. The method is based on an aggregation principle which is well established in the literature and features a multiplicative coarse-level correction. Recursive application of the aggregation principle, which uses an operator-dependent coarsening, yields a multi-level method which has been shown experimentally to give results significantly faster than the typical methods currently in use. When cast as a multigrid-like method, the algorithm is seen to be a Galerkin-Full Approximation Scheme with a solution-dependent prolongation operator. Special properties of this prolongation lead to the cancellation of the computationally intensive terms of the coarse-level equations.
On the multi-level solution algorithm for Markov chains
Horton, G.
1996-12-31
We discuss the recently introduced multi-level algorithm for the steady-state solution of Markov chains. The method is based on the aggregation principle, which is well established in the literature. Recursive application of the aggregation yields a multi-level method which has been shown experimentally to give results significantly faster than the methods currently in use. The algorithm can be reformulated as an algebraic multigrid scheme of Galerkin-full approximation type. The uniqueness of the scheme stems from its solution-dependent prolongation operator which permits significant computational savings in the evaluation of certain terms. This paper describes the modeling of computer systems to derive information on performance, measured typically as job throughput or component utilization, and availability, defined as the proportion of time a system is able to perform a certain function in the presence of component failures and possibly also repairs.
An algorithm for the solution of dynamic linear programs
NASA Technical Reports Server (NTRS)
Psiaki, Mark L.
1989-01-01
The algorithm's objective is to efficiently solve Dynamic Linear Programs (DLP) by taking advantage of their special staircase structure. This algorithm constitutes a stepping stone to an improved algorithm for solving Dynamic Quadratic Programs, which, in turn, would make the nonlinear programming method of Successive Quadratic Programs more practical for solving trajectory optimization problems. The ultimate goal is to being trajectory optimization solution speeds into the realm of real-time control. The algorithm exploits the staircase nature of the large constraint matrix of the equality-constrained DLPs encountered when solving inequality-constrained DLPs by an active set approach. A numerically-stable, staircase QL factorization of the staircase constraint matrix is carried out starting from its last rows and columns. The resulting recursion is like the time-varying Riccati equation from multi-stage LQR theory. The resulting factorization increases the efficiency of all of the typical LP solution operations over that of a dense matrix LP code. At the same time numerical stability is ensured. The algorithm also takes advantage of dynamic programming ideas about the cost-to-go by relaxing active pseudo constraints in a backwards sweeping process. This further decreases the cost per update of the LP rank-1 updating procedure, although it may result in more changes of the active set that if pseudo constraints were relaxed in a non-stagewise fashion. The usual stability of closed-loop Linear/Quadratic optimally-controlled systems, if it carries over to strictly linear cost functions, implies that the saving due to reduced factor update effort may outweigh the cost of an increased number of updates. An aerospace example is presented in which a ground-to-ground rocket's distance is maximized. This example demonstrates the applicability of this class of algorithms to aerospace guidance. It also sheds light on the efficacy of the proposed pseudo constraint relaxation
A new mathematical adjoint for the modified SAAF_{-SN} equations
Schunert, Sebastian; Wang, Yaqi; Martineau, Richard; DeHart, Mark D.
2015-01-01
We present a new adjoint FEM weak form, which can be directly used for evaluating the mathematical adjoint, suitable for perturbation calculations, of the self-adjoint angular flux SN equations (SAAF_{-SN}) without construction and transposition of the underlying coefficient matrix. Stabilization schemes incorporated in the described SAAF_{-SN} method make the mathematical adjoint distinct from the physical adjoint, i.e. the solution of the continuous adjoint equation with SAAF_{-SN} . This weak form is implemented into RattleSnake, the MOOSE (Multiphysics Object-Oriented Simulation Environment) based transport solver. Numerical results verify the correctness of the implementation and show its utility both for fixed source and eigenvalue problems.
Reentry-Vehicle Shape Optimization Using a Cartesian Adjoint Method and CAD Geometry
NASA Technical Reports Server (NTRS)
Nemec, Marian; Aftosmis, Michael J.
2006-01-01
A DJOINT solutions of the governing flow equations are becoming increasingly important for the development of efficient analysis and optimization algorithms. A well-known use of the adjoint method is gradient-based shape. Given an objective function that defines some measure of performance, such as the lift and drag functionals, its gradient is computed at a cost that is essentially independent of the number of design variables (e.g., geometric parameters that control the shape). Classic aerodynamic applications of gradient-based optimization include the design of cruise configurations for transonic and supersonic flow, as well as the design of high-lift systems. are perhaps the most promising approach for addressing the issues of flow solution automation for aerodynamic design problems. In these methods, the discretization of the wetted surface is decoupled from that of the volume mesh. This not only enables fast and robust mesh generation for geometry of arbitrary complexity, but also facilitates access to geometry modeling and manipulation using parametric computer-aided design (CAD). In previous work on Cartesian adjoint solvers, Melvin et al. developed an adjoint formulation for the TRANAIR code, which is based on the full-potential equation with viscous corrections. More recently, Dadone and Grossman presented an adjoint formulation for the two-dimensional Euler equations using a ghost-cell method to enforce the wall boundary conditions. In Refs. 18 and 19, we presented an accurate and efficient algorithm for the solution of the adjoint Euler equations discretized on Cartesian meshes with embedded, cut-cell boundaries. Novel aspects of the algorithm were the computation of surface shape sensitivities for triangulations based on parametric-CAD models and the linearization of the coupling between the surface triangulation and the cut-cells. The accuracy of the gradient computation was verified using several three-dimensional test cases, which included design
Multigrid methods for bifurcation problems: The self adjoint case
NASA Technical Reports Server (NTRS)
Taasan, Shlomo
1987-01-01
This paper deals with multigrid methods for computational problems that arise in the theory of bifurcation and is restricted to the self adjoint case. The basic problem is to solve for arcs of solutions, a task that is done successfully with an arc length continuation method. Other important issues are, for example, detecting and locating singular points as part of the continuation process, switching branches at bifurcation points, etc. Multigrid methods have been applied to continuation problems. These methods work well at regular points and at limit points, while they may encounter difficulties in the vicinity of bifurcation points. A new continuation method that is very efficient also near bifurcation points is presented here. The other issues mentioned above are also treated very efficiently with appropriate multigrid algorithms. For example, it is shown that limit points and bifurcation points can be solved for directly by a multigrid algorithm. Moreover, the algorithms presented here solve the corresponding problems in just a few work units (about 10 or less), where a work unit is the work involved in one local relaxation on the finest grid.
An algorithm for the numerical solution of linear differential games
Polovinkin, E S; Ivanov, G E; Balashov, M V; Konstantinov, R V; Khorev, A V
2001-10-31
A numerical algorithm for the construction of stable Krasovskii bridges, Pontryagin alternating sets, and also of piecewise program strategies solving two-person linear differential (pursuit or evasion) games on a fixed time interval is developed on the basis of a general theory. The aim of the first player (the pursuer) is to hit a prescribed target (terminal) set by the phase vector of the control system at the prescribed time. The aim of the second player (the evader) is the opposite. A description of numerical algorithms used in the solution of differential games of the type under consideration is presented and estimates of the errors resulting from the approximation of the game sets by polyhedra are presented.
Learning a trajectory using adjoint functions and teacher forcing
NASA Technical Reports Server (NTRS)
Toomarian, Nikzad B.; Barhen, Jacob
1992-01-01
A new methodology for faster supervised temporal learning in nonlinear neural networks is presented which builds upon the concept of adjoint operators to allow fast computation of the gradients of an error functional with respect to all parameters of the neural architecture, and exploits the concept of teacher forcing to incorporate information on the desired output into the activation dynamics. The importance of the initial or final time conditions for the adjoint equations is discussed. A new algorithm is presented in which the adjoint equations are solved simultaneously (i.e., forward in time) with the activation dynamics of the neural network. We also indicate how teacher forcing can be modulated in time as learning proceeds. The results obtained show that the learning time is reduced by one to two orders of magnitude with respect to previously published results, while trajectory tracking is significantly improved. The proposed methodology makes hardware implementation of temporal learning attractive for real-time applications.
Self-adaptive closed constrained solution algorithms for nonlinear conduction
NASA Technical Reports Server (NTRS)
Padovan, J.; Tovichakchaikul, S.
1982-01-01
Self-adaptive solution algorithms are developed for nonlinear heat conduction problems encountered in analyzing materials for use in high temperature or cryogenic conditions. The nonlinear effects are noted to occur due to convection and radiation effects, as well as temperature-dependent properties of the materials. Incremental successive substitution (ISS) and Newton-Raphson (NR) procedures are treated as extrapolation schemes which have solution projections bounded by a hyperline with an externally applied thermal load vector arising from internal heat generation and boundary conditions. Closed constraints are formulated which improve the efficiency and stability of the procedures by employing closed ellipsoidal surfaces to control the size of successive iterations. Governing equations are defined for nonlinear finite element models, and comparisons are made of results using the the new method and the ISS and NR schemes for epoxy, PVC, and CuGe.
NASA Astrophysics Data System (ADS)
Bodin, Jacques
2015-03-01
In this study, new multi-dimensional time-domain random walk (TDRW) algorithms are derived from approximate one-dimensional (1-D), two-dimensional (2-D), and three-dimensional (3-D) analytical solutions of the advection-dispersion equation and from exact 1-D, 2-D, and 3-D analytical solutions of the pure-diffusion equation. These algorithms enable the calculation of both the time required for a particle to travel a specified distance in a homogeneous medium and the mass recovery at the observation point, which may be incomplete due to 2-D or 3-D transverse dispersion or diffusion. The method is extended to heterogeneous media, represented as a piecewise collection of homogeneous media. The particle motion is then decomposed along a series of intermediate checkpoints located on the medium interface boundaries. The accuracy of the multi-dimensional TDRW method is verified against (i) exact analytical solutions of solute transport in homogeneous media and (ii) finite-difference simulations in a synthetic 2-D heterogeneous medium of simple geometry. The results demonstrate that the method is ideally suited to purely diffusive transport and to advection-dispersion transport problems dominated by advection. Conversely, the method is not recommended for highly dispersive transport problems because the accuracy of the advection-dispersion TDRW algorithms degrades rapidly for a low Péclet number, consistent with the accuracy limit of the approximate analytical solutions. The proposed approach provides a unified methodology for deriving multi-dimensional time-domain particle equations and may be applicable to other mathematical transport models, provided that appropriate analytical solutions are available.
Heinstein, M.W.
1997-10-01
A contact enforcement algorithm has been developed for matrix-free quasistatic finite element techniques. Matrix-free (iterative) solution algorithms such as nonlinear Conjugate Gradients (CG) and Dynamic Relaxation (DR) are distinctive in that the number of iterations required for convergence is typically of the same order as the number of degrees of freedom of the model. From iteration to iteration the contact normal and tangential forces vary significantly making contact constraint satisfaction tenuous. Furthermore, global determination and enforcement of the contact constraints every iteration could be questioned on the grounds of efficiency. This work addresses this situation by introducing an intermediate iteration for treating the active gap constraint and at the same time exactly (kinematically) enforcing the linearized gap rate constraint for both frictionless and frictional response.
A Posteriori Analysis for Hydrodynamic Simulations Using Adjoint Methodologies
Woodward, C S; Estep, D; Sandelin, J; Wang, H
2009-02-26
This report contains results of analysis done during an FY08 feasibility study investigating the use of adjoint methodologies for a posteriori error estimation for hydrodynamics simulations. We developed an approach to adjoint analysis for these systems through use of modified equations and viscosity solutions. Targeting first the 1D Burgers equation, we include a verification of the adjoint operator for the modified equation for the Lax-Friedrichs scheme, then derivations of an a posteriori error analysis for a finite difference scheme and a discontinuous Galerkin scheme applied to this problem. We include some numerical results showing the use of the error estimate. Lastly, we develop a computable a posteriori error estimate for the MAC scheme applied to stationary Navier-Stokes.
The development of solution algorithms for compressible flows
NASA Astrophysics Data System (ADS)
Slack, David Christopher
Three main topics were examined. The first is the development and comparison of time integration schemes on 2-D unstructured meshes. Both explicit and implicit solution grids are presented. Cell centered and cell vertex finite volume upwind schemes using Roe's approximate Riemann solver are developed. The second topic involves an interactive adaptive remeshing algorithm which uses a frontal grid generator and is compared to a single grid calculation. The final topic examined is the capabilities developed for a structured 3-D code called GASP. The capabilities include: generalized chemistry and thermodynamic modeling, space marching, memory management through the use of binary C I/O, and algebraic and two equation eddy viscosity turbulence modeling. Results are given for Mach 1.7 3-D analytic forebody, a Mach 1.38 axisymmetric nozzle with hydrogen-air combustion, a Mach 14.15 deg ramp, and Mach 0.3 viscous flow over a flat plate.
Adjoint Sensitivity Computations for an Embedded-Boundary Cartesian Mesh Method and CAD Geometry
NASA Technical Reports Server (NTRS)
Nemec, Marian; Aftosmis,Michael J.
2006-01-01
Cartesian-mesh methods are perhaps the most promising approach for addressing the issues of flow solution automation for aerodynamic design problems. In these methods, the discretization of the wetted surface is decoupled from that of the volume mesh. This not only enables fast and robust mesh generation for geometry of arbitrary complexity, but also facilitates access to geometry modeling and manipulation using parametric Computer-Aided Design (CAD) tools. Our goal is to combine the automation capabilities of Cartesian methods with an eficient computation of design sensitivities. We address this issue using the adjoint method, where the computational cost of the design sensitivities, or objective function gradients, is esseutially indepeudent of the number of design variables. In previous work, we presented an accurate and efficient algorithm for the solution of the adjoint Euler equations discretized on Cartesian meshes with embedded, cut-cell boundaries. Novel aspects of the algorithm included the computation of surface shape sensitivities for triangulations based on parametric-CAD models and the linearization of the coupling between the surface triangulation and the cut-cells. The objective of the present work is to extend our adjoint formulation to problems involving general shape changes. Central to this development is the computation of volume-mesh sensitivities to obtain a reliable approximation of the objective finction gradient. Motivated by the success of mesh-perturbation schemes commonly used in body-fitted unstructured formulations, we propose an approach based on a local linearization of a mesh-perturbation scheme similar to the spring analogy. This approach circumvents most of the difficulties that arise due to non-smooth changes in the cut-cell layer as the boundary shape evolves and provides a consistent approximation tot he exact gradient of the discretized abjective function. A detailed gradient accurace study is presented to verify our approach
NASA Technical Reports Server (NTRS)
Nguyen, Nhan T.; Hornby, Gregory; Ishihara, Abe
2013-01-01
This paper describes two methods of trajectory optimization to obtain an optimal trajectory of minimum-fuel- to-climb for an aircraft. The first method is based on the adjoint method, and the second method is based on a direct trajectory optimization method using a Chebyshev polynomial approximation and cubic spine approximation. The approximate optimal trajectory will be compared with the adjoint-based optimal trajectory which is considered as the true optimal solution of the trajectory optimization problem. The adjoint-based optimization problem leads to a singular optimal control solution which results in a bang-singular-bang optimal control.
Applicability of the Newman-Janis algorithm to black hole solutions of modified gravity theories
NASA Astrophysics Data System (ADS)
Hansen, Devin; Yunes, Nicolás
2013-11-01
The Newman-Janis algorithm has been widely used to construct rotating black hole solutions from nonrotating counterparts. While this algorithm was developed within general relativity (GR), it has more recently been applied to nonrotating solutions in modified gravity theories. We find that the application of the Newman-Janis algorithm to an arbitrary non-GR spherically symmetric solution introduces pathologies in the resulting axially symmetric metric. This then establishes that, in general, the Newman-Janis algorithm should not be used to construct rotating black hole solutions outside of General Relativity.
NASA Technical Reports Server (NTRS)
Reuther, James; Jameson, Antony; Alonso, Juan Jose; Rimlinger, Mark J.; Saunders, David
1997-01-01
An aerodynamic shape optimization method that treats the design of complex aircraft configurations subject to high fidelity computational fluid dynamics (CFD), geometric constraints and multiple design points is described. The design process will be greatly accelerated through the use of both control theory and distributed memory computer architectures. Control theory is employed to derive the adjoint differential equations whose solution allows for the evaluation of design gradient information at a fraction of the computational cost required by previous design methods. The resulting problem is implemented on parallel distributed memory architectures using a domain decomposition approach, an optimized communication schedule, and the MPI (Message Passing Interface) standard for portability and efficiency. The final result achieves very rapid aerodynamic design based on a higher order CFD method. In order to facilitate the integration of these high fidelity CFD approaches into future multi-disciplinary optimization (NW) applications, new methods must be developed which are capable of simultaneously addressing complex geometries, multiple objective functions, and geometric design constraints. In our earlier studies, we coupled the adjoint based design formulations with unconstrained optimization algorithms and showed that the approach was effective for the aerodynamic design of airfoils, wings, wing-bodies, and complex aircraft configurations. In many of the results presented in these earlier works, geometric constraints were satisfied either by a projection into feasible space or by posing the design space parameterization such that it automatically satisfied constraints. Furthermore, with the exception of reference 9 where the second author initially explored the use of multipoint design in conjunction with adjoint formulations, our earlier works have focused on single point design efforts. Here we demonstrate that the same methodology may be extended to treat
ANISORROPIA: the adjoint of the aerosol thermodynamic model ISORROPIA
NASA Astrophysics Data System (ADS)
Capps, S. L.; Henze, D. K.; Hakami, A.; Russell, A. G.; Nenes, A.
2011-08-01
We present the development of ANISORROPIA, the discrete adjoint of the ISORROPIA thermodynamic equilibrium model that treats the Na+-SO42--HSO4--NH4+-NO3--Cl--H2O aerosol system, and we demonstrate its sensitivity analysis capabilities. ANISORROPIA calculates sensitivities of an inorganic species in aerosol or gas phase with respect to the total concentrations of each species present with only a two-fold increase in computational time over the forward model execution. Due to the highly nonlinear and discontinuous solution surface of ISORROPIA, evaluation of the adjoint required a new, complex-variable version of the the model, which determines first-order sensitivities with machine precision and avoids cancellation errors arising from finite difference calculations. The adjoint is verified over an atmospherically relevant range of concentrations, temperature, and relative humidity. We apply ANISORROPIA to recent field campaign results from Atlanta, GA, USA, and Mexico City, Mexico, to characterize the inorganic aerosol sensitivities of these distinct urban air masses. The variability in the relationship between PM2.5 mass and precursor concentrations shown has important implications for air quality and climate. ANISORROPIA enables efficient elucidation of aerosol concentration dependence on aerosol precursor emissions in the context of atmospheric chemical transport model adjoints.
A fast algorithm for numerical solutions to Fortet's equation
NASA Astrophysics Data System (ADS)
Brumen, Gorazd
2008-10-01
A fast algorithm for computation of default times of multiple firms in a structural model is presented. The algorithm uses a multivariate extension of the Fortet's equation and the structure of Toeplitz matrices to significantly improve the computation time. In a financial market consisting of M[not double greater-than sign]1 firms and N discretization points in every dimension the algorithm uses O(nlogn·M·M!·NM(M-1)/2) operations, where n is the number of discretization points in the time domain. The algorithm is applied to firm survival probability computation and zero coupon bond pricing.
Self-adjointness of deformed unbounded operators
Much, Albert
2015-09-15
We consider deformations of unbounded operators by using the novel construction tool of warped convolutions. By using the Kato-Rellich theorem, we show that unbounded self-adjoint deformed operators are self-adjoint if they satisfy a certain condition. This condition proves itself to be necessary for the oscillatory integral to be well-defined. Moreover, different proofs are given for self-adjointness of deformed unbounded operators in the context of quantum mechanics and quantum field theory.
Application of Adjoint Methodology in Various Aspects of Sonic Boom Design
NASA Technical Reports Server (NTRS)
Rallabhandi, Sriram K.
2014-01-01
One of the advances in computational design has been the development of adjoint methods allowing efficient calculation of sensitivities in gradient-based shape optimization. This paper discusses two new applications of adjoint methodology that have been developed to aid in sonic boom mitigation exercises. In the first, equivalent area targets are generated using adjoint sensitivities of selected boom metrics. These targets may then be used to drive the vehicle shape during optimization. The second application is the computation of adjoint sensitivities of boom metrics on the ground with respect to parameters such as flight conditions, propagation sampling rate, and selected inputs to the propagation algorithms. These sensitivities enable the designer to make more informed selections of flight conditions at which the chosen cost functionals are less sensitive.
Fast Combinatorial Algorithm for the Solution of Linearly Constrained Least Squares Problems
Van Benthem, Mark H.; Keenan, Michael R.
2008-11-11
A fast combinatorial algorithm can significantly reduce the computational burden when solving general equality and inequality constrained least squares problems with large numbers of observation vectors. The combinatorial algorithm provides a mathematically rigorous solution and operates at great speed by reorganizing the calculations to take advantage of the combinatorial nature of the problems to be solved. The combinatorial algorithm exploits the structure that exists in large-scale problems in order to minimize the number of arithmetic operations required to obtain a solution.
Numerical Computation of Sensitivities and the Adjoint Approach
NASA Technical Reports Server (NTRS)
Lewis, Robert Michael
1997-01-01
We discuss the numerical computation of sensitivities via the adjoint approach in optimization problems governed by differential equations. We focus on the adjoint problem in its weak form. We show how one can avoid some of the problems with the adjoint approach, such as deriving suitable boundary conditions for the adjoint equation. We discuss the convergence of numerical approximations of the costate computed via the weak form of the adjoint problem and show the significance for the discrete adjoint problem.
Evolutionary Algorithms Approach to the Solution of Damage Detection Problems
NASA Astrophysics Data System (ADS)
Salazar Pinto, Pedro Yoajim; Begambre, Oscar
2010-09-01
In this work is proposed a new Self-Configured Hybrid Algorithm by combining the Particle Swarm Optimization (PSO) and a Genetic Algorithm (GA). The aim of the proposed strategy is to increase the stability and accuracy of the search. The central idea is the concept of Guide Particle, this particle (the best PSO global in each generation) transmits its information to a particle of the following PSO generation, which is controlled by the GA. Thus, the proposed hybrid has an elitism feature that improves its performance and guarantees the convergence of the procedure. In different test carried out in benchmark functions, reported in the international literature, a better performance in stability and accuracy was observed; therefore the new algorithm was used to identify damage in a simple supported beam using modal data. Finally, it is worth noting that the algorithm is independent of the initial definition of heuristic parameters.
Adjoint calculations for multiple scattering of Compton and Rayleigh effects
NASA Astrophysics Data System (ADS)
Fernández, J. E.; Sumini, M.
1992-08-01
As is well known, the experimental determination of the Compton profile requires a particular geometry with a scattering angle close to π. That situation involves a narrow multiple-scattering spectrum that overlaps the Compton peak, making it difficult to analyze the different contributions to the profile. We show how the solution of the adjoint problem can help in devising more useful experimental configurations, giving, through its classical "importance" meaning, a formally clear picture of the whole problem.
ASYMPTOTICALLY OPTIMAL HIGH-ORDER ACCURATE ALGORITHMS FOR THE SOLUTION OF CERTAIN ELLIPTIC PDEs
Leonid Kunyansky, PhD
2008-11-26
The main goal of the project, "Asymptotically Optimal, High-Order Accurate Algorithms for the Solution of Certain Elliptic PDE's" (DE-FG02-03ER25577) was to develop fast, high-order algorithms for the solution of scattering problems and spectral problems of photonic crystals theory. The results we obtained lie in three areas: (1) asymptotically fast, high-order algorithms for the solution of eigenvalue problems of photonics, (2) fast, high-order algorithms for the solution of acoustic and electromagnetic scattering problems in the inhomogeneous media, and (3) inversion formulas and fast algorithms for the inverse source problem for the acoustic wave equation, with applications to thermo- and opto- acoustic tomography.
A Numerical Algorithm for the Solution of a Phase-Field Model of Polycrystalline Materials
Dorr, M R; Fattebert, J; Wickett, M E; Belak, J F; Turchi, P A
2008-12-04
We describe an algorithm for the numerical solution of a phase-field model (PFM) of microstructure evolution in polycrystalline materials. The PFM system of equations includes a local order parameter, a quaternion representation of local orientation and a species composition parameter. The algorithm is based on the implicit integration of a semidiscretization of the PFM system using a backward difference formula (BDF) temporal discretization combined with a Newton-Krylov algorithm to solve the nonlinear system at each time step. The BDF algorithm is combined with a coordinate projection method to maintain quaternion unit length, which is related to an important solution invariant. A key element of the Newton-Krylov algorithm is the selection of a preconditioner to accelerate the convergence of the Generalized Minimum Residual algorithm used to solve the Jacobian linear system in each Newton step. Results are presented for the application of the algorithm to 2D and 3D examples.
A multi-level solution algorithm for steady-state Markov chains
NASA Technical Reports Server (NTRS)
Horton, Graham; Leutenegger, Scott T.
1993-01-01
A new iterative algorithm, the multi-level algorithm, for the numerical solution of steady state Markov chains is presented. The method utilizes a set of recursively coarsened representations of the original system to achieve accelerated convergence. It is motivated by multigrid methods, which are widely used for fast solution of partial differential equations. Initial results of numerical experiments are reported, showing significant reductions in computation time, often an order of magnitude or more, relative to the Gauss-Seidel and optimal SOR algorithms for a variety of test problems. The multi-level method is compared and contrasted with the iterative aggregation-disaggregation algorithm of Takahashi.
Fast algorithm for the solution of large-scale non-negativity constrained least squares problems.
Van Benthem, Mark Hilary; Keenan, Michael Robert
2004-06-01
Algorithms for multivariate image analysis and other large-scale applications of multivariate curve resolution (MCR) typically employ constrained alternating least squares (ALS) procedures in their solution. The solution to a least squares problem under general linear equality and inequality constraints can be reduced to the solution of a non-negativity-constrained least squares (NNLS) problem. Thus the efficiency of the solution to any constrained least square problem rests heavily on the underlying NNLS algorithm. We present a new NNLS solution algorithm that is appropriate to large-scale MCR and other ALS applications. Our new algorithm rearranges the calculations in the standard active set NNLS method on the basis of combinatorial reasoning. This rearrangement serves to reduce substantially the computational burden required for NNLS problems having large numbers of observation vectors.
An algorithm for the systematic disturbance of optimal rotational solutions
NASA Technical Reports Server (NTRS)
Grunwald, Arthur J.; Kaiser, Mary K.
1989-01-01
An algorithm for introducing a systematic rotational disturbance into an optimal (i.e., single axis) rotational trajectory is described. This disturbance introduces a motion vector orthogonal to the quaternion-defined optimal rotation axis. By altering the magnitude of this vector, the degree of non-optimality can be controlled. The metric properties of the distortion parameter are described, with analogies to two-dimensional translational motion. This algorithm was implemented in a motion-control program on a three-dimensional graphic workstation. It supports a series of human performance studies on the detectability of rotational trajectory optimality by naive observers.
Using Strassen's algorithm to accelerate the solution of linear systems
NASA Technical Reports Server (NTRS)
Bailey, David H.; Lee, King; Simon, Horst D.
1990-01-01
Strassen's algorithm for fast matrix-matrix multiplication has been implemented for matrices of arbitrary shapes on the CRAY-2 and CRAY Y-MP supercomputers. Several techniques have been used to reduce the scratch space requirement for this algorithm while simultaneously preserving a high level of performance. When the resulting Strassen-based matrix multiply routine is combined with some routines from the new LAPACK library, LU decomposition can be performed with rates significantly higher than those achieved by conventional means. We succeeded in factoring a 2048 x 2048 matrix on the CRAY Y-MP at a rate equivalent to 325 MFLOPS.
NASA Technical Reports Server (NTRS)
Yamaleev, N. K.; Diskin, B.; Nielsen, E. J.
2009-01-01
.We study local-in-time adjoint-based methods for minimization of ow matching functionals subject to the 2-D unsteady compressible Euler equations. The key idea of the local-in-time method is to construct a very accurate approximation of the global-in-time adjoint equations and the corresponding sensitivity derivative by using only local information available on each time subinterval. In contrast to conventional time-dependent adjoint-based optimization methods which require backward-in-time integration of the adjoint equations over the entire time interval, the local-in-time method solves local adjoint equations sequentially over each time subinterval. Since each subinterval contains relatively few time steps, the storage cost of the local-in-time method is much lower than that of the global adjoint formulation, thus making the time-dependent optimization feasible for practical applications. The paper presents a detailed comparison of the local- and global-in-time adjoint-based methods for minimization of a tracking functional governed by the Euler equations describing the ow around a circular bump. Our numerical results show that the local-in-time method converges to the same optimal solution obtained with the global counterpart, while drastically reducing the memory cost as compared to the global-in-time adjoint formulation.
Neural Networks Art: Solving Problems with Multiple Solutions and New Teaching Algorithm
Dmitrienko, V. D; Zakovorotnyi, A. Yu.; Leonov, S. Yu.; Khavina, I. P
2014-01-01
A new discrete neural networks adaptive resonance theory (ART), which allows solving problems with multiple solutions, is developed. New algorithms neural networks teaching ART to prevent degradation and reproduction classes at training noisy input data is developed. Proposed learning algorithms discrete ART networks, allowing obtaining different classification methods of input. PMID:25246988
Yurtkuran, Alkın
2016-01-01
The artificial bee colony (ABC) algorithm is a popular swarm based technique, which is inspired from the intelligent foraging behavior of honeybee swarms. This paper proposes a new variant of ABC algorithm, namely, enhanced ABC with solution acceptance rule and probabilistic multisearch (ABC-SA) to address global optimization problems. A new solution acceptance rule is proposed where, instead of greedy selection between old solution and new candidate solution, worse candidate solutions have a probability to be accepted. Additionally, the acceptance probability of worse candidates is nonlinearly decreased throughout the search process adaptively. Moreover, in order to improve the performance of the ABC and balance the intensification and diversification, a probabilistic multisearch strategy is presented. Three different search equations with distinctive characters are employed using predetermined search probabilities. By implementing a new solution acceptance rule and a probabilistic multisearch approach, the intensification and diversification performance of the ABC algorithm is improved. The proposed algorithm has been tested on well-known benchmark functions of varying dimensions by comparing against novel ABC variants, as well as several recent state-of-the-art algorithms. Computational results show that the proposed ABC-SA outperforms other ABC variants and is superior to state-of-the-art algorithms proposed in the literature. PMID:26819591
Yurtkuran, Alkın; Emel, Erdal
2016-01-01
The artificial bee colony (ABC) algorithm is a popular swarm based technique, which is inspired from the intelligent foraging behavior of honeybee swarms. This paper proposes a new variant of ABC algorithm, namely, enhanced ABC with solution acceptance rule and probabilistic multisearch (ABC-SA) to address global optimization problems. A new solution acceptance rule is proposed where, instead of greedy selection between old solution and new candidate solution, worse candidate solutions have a probability to be accepted. Additionally, the acceptance probability of worse candidates is nonlinearly decreased throughout the search process adaptively. Moreover, in order to improve the performance of the ABC and balance the intensification and diversification, a probabilistic multisearch strategy is presented. Three different search equations with distinctive characters are employed using predetermined search probabilities. By implementing a new solution acceptance rule and a probabilistic multisearch approach, the intensification and diversification performance of the ABC algorithm is improved. The proposed algorithm has been tested on well-known benchmark functions of varying dimensions by comparing against novel ABC variants, as well as several recent state-of-the-art algorithms. Computational results show that the proposed ABC-SA outperforms other ABC variants and is superior to state-of-the-art algorithms proposed in the literature. PMID:26819591
Consistent Adjoint Driven Importance Sampling using Space, Energy and Angle
Peplow, Douglas E.; Mosher, Scott W; Evans, Thomas M
2012-08-01
For challenging radiation transport problems, hybrid methods combine the accuracy of Monte Carlo methods with the global information present in deterministic methods. One of the most successful hybrid methods is CADIS Consistent Adjoint Driven Importance Sampling. This method uses a deterministic adjoint solution to construct a biased source distribution and consistent weight windows to optimize a specific tally in a Monte Carlo calculation. The method has been implemented into transport codes using just the spatial and energy information from the deterministic adjoint and has been used in many applications to compute tallies with much higher figures-of-merit than analog calculations. CADIS also outperforms user-supplied importance values, which usually take long periods of user time to develop. This work extends CADIS to develop weight windows that are a function of the position, energy, and direction of the Monte Carlo particle. Two types of consistent source biasing are presented: one method that biases the source in space and energy while preserving the original directional distribution and one method that biases the source in space, energy, and direction. Seven simple example problems are presented which compare the use of the standard space/energy CADIS with the new space/energy/angle treatments.
ANISORROPIA: the adjoint of the aerosol thermodynamic model ISORROPIA
NASA Astrophysics Data System (ADS)
Capps, S. L.; Henze, D. K.; Hakami, A.; Russell, A. G.; Nenes, A.
2012-01-01
We present the development of ANISORROPIA, the discrete adjoint of the ISORROPIA thermodynamic equilibrium model that treats the Na+-SO42-- HSO4--NH4+ -NO3--Cl--H2O aerosol system, and we demonstrate its sensitivity analysis capabilities. ANISORROPIA calculates sensitivities of an inorganic species in aerosol or gas phase with respect to the total concentrations of each species present with less than a two-fold increase in computational time over the concentration calculations. Due to the highly nonlinear and discontinuous solution surface of ISORROPIA, evaluation of the adjoint required a new, complex-variable version of the model, which determines first-order sensitivities with machine precision and avoids cancellation errors arising from finite difference calculations. The adjoint is verified over an atmospherically relevant range of concentrations, temperature, and relative humidity. We apply ANISORROPIA to recent field campaign results from Atlanta, GA, USA, and Mexico City, Mexico, to characterize the inorganic aerosol sensitivities of these distinct urban air masses. The variability in the relationship between fine mode inorganic aerosol mass and precursor concentrations shown has important implications for air quality and climate.
An Efficient Algorithm for Partitioning and Authenticating Problem-Solutions of eLeaming Contents
ERIC Educational Resources Information Center
Dewan, Jahangir; Chowdhury, Morshed; Batten, Lynn
2013-01-01
Content authenticity and correctness is one of the important challenges in eLearning as there can be many solutions to one specific problem in cyber space. Therefore, the authors feel it is necessary to map problems to solutions using graph partition and weighted bipartite matching. This article proposes an efficient algorithm to partition…
ELASTIC NET FOR COX'S PROPORTIONAL HAZARDS MODEL WITH A SOLUTION PATH ALGORITHM.
Wu, Yichao
2012-01-01
For least squares regression, Efron et al. (2004) proposed an efficient solution path algorithm, the least angle regression (LAR). They showed that a slight modification of the LAR leads to the whole LASSO solution path. Both the LAR and LASSO solution paths are piecewise linear. Recently Wu (2011) extended the LAR to generalized linear models and the quasi-likelihood method. In this work we extend the LAR further to handle Cox's proportional hazards model. The goal is to develop a solution path algorithm for the elastic net penalty (Zou and Hastie (2005)) in Cox's proportional hazards model. This goal is achieved in two steps. First we extend the LAR to optimizing the log partial likelihood plus a fixed small ridge term. Then we define a path modification, which leads to the solution path of the elastic net regularized log partial likelihood. Our solution path is exact and piecewise determined by ordinary differential equation systems. PMID:23226932
Comparative study of fusion algorithms and implementation of new efficient solution
NASA Astrophysics Data System (ADS)
Besrour, Amine; Snoussi, Hichem; Siala, Mohamed; Abdelkefi, Fatma
2014-05-01
High Dynamic Range (HDR) imaging has been the subject of significant researches over the past years, the goal of acquiring the best cinema-quality HDR images of fast-moving scenes using an efficient merging algorithm has not yet been achieved. In fact, through the years, many efficient algorithms have been implemented and developed. However, they are not yet efficient since they don't treat all the situations and they have not enough speed to ensure fast HDR image reconstitution. In this paper, we will present a full comparative analyze and study of the available fusion algorithms. Also, we will implement our personal algorithm which can be more optimized and faster than the existed ones. We will also present our investigated algorithm that has the advantage to be more optimized than the existing ones. This merging algorithm is related to our hardware solution allowing us to obtain four pictures with different exposures.
A new algorithm for generating highly accurate benchmark solutions to transport test problems
Azmy, Y.Y.
1997-06-01
We present a new algorithm for solving the neutron transport equation in its discrete-variable form. The new algorithm is based on computing the full matrix relating the scalar flux spatial moments in all cells to the fixed neutron source spatial moments, foregoing the need to compute the angular flux spatial moments, and thereby eliminating the need for sweeping the spatial mesh in each discrete-angular direction. The matrix equation is solved exactly in test cases, producing a solution vector that is free from iteration convergence error, and subject only to truncation and roundoff errors. Our algorithm is designed to provide method developers with a quick and simple solution scheme to test their new methods on difficult test problems without the need to develop sophisticated solution techniques, e.g. acceleration, before establishing the worthiness of their innovation. We demonstrate the utility of the new algorithm by applying it to the Arbitrarily High Order Transport Nodal (AHOT-N) method, and using it to solve two of Burre`s Suite of Test Problems (BSTP). Our results provide highly accurate benchmark solutions, that can be distributed electronically and used to verify the pointwise accuracy of other solution methods and algorithms.
NEMOTAM: tangent and adjoint models for the ocean modelling platform NEMO
NASA Astrophysics Data System (ADS)
Vidard, A.; Bouttier, P.-A.; Vigilant, F.
2014-10-01
The tangent linear and adjoint model (TAM) are efficient tools to analyse and to control dynamical systems such as NEMO. They can be involved in a large range of applications such as sensitivity analysis, parameter estimation or the computation of characteristics vectors. TAM is also required by the 4-D-VAR algorithm which is one of the major method in Data Assimilation. This paper describes the development and the validation of the Tangent linear and Adjoint Model for the NEMO ocean modelling platform (NEMOTAM). The diagnostic tools that are available alongside NEMOTAM are detailed and discussed and several applications are also presented.
Solution algorithms for the two-dimensional Euler equations on unstructured meshes
NASA Technical Reports Server (NTRS)
Whitaker, D. L.; Slack, David C.; Walters, Robert W.
1990-01-01
The objective of the study was to analyze implicit techniques employed in structured grid algorithms for solving two-dimensional Euler equations and extend them to unstructured solvers in order to accelerate convergence rates. A comparison is made between nine different algorithms for both first-order and second-order accurate solutions. Higher-order accuracy is achieved by using multidimensional monotone linear reconstruction procedures. The discussion is illustrated by results for flow over a transonic circular arc.
Finite element solution for energy conservation using a highly stable explicit integration algorithm
NASA Technical Reports Server (NTRS)
Baker, A. J.; Manhardt, P. D.
1972-01-01
Theoretical derivation of a finite element solution algorithm for the transient energy conservation equation in multidimensional, stationary multi-media continua with irregular solution domain closure is considered. The complete finite element matrix forms for arbitrarily irregular discretizations are established, using natural coordinate function representations. The algorithm is embodied into a user-oriented computer program (COMOC) which obtains transient temperature distributions at the node points of the finite element discretization using a highly stable explicit integration procedure with automatic error control features. The finite element algorithm is shown to posses convergence with discretization for a transient sample problem. The condensed form for the specific heat element matrix is shown to be preferable to the consistent form. Computed results for diverse problems illustrate the versatility of COMOC, and easily prepared output subroutines are shown to allow quick engineering assessment of solution behavior.
A practical discrete-adjoint method for high-fidelity compressible turbulence simulations
NASA Astrophysics Data System (ADS)
Vishnampet, Ramanathan; Bodony, Daniel J.; Freund, Jonathan B.
2015-03-01
Methods and computing hardware advances have enabled accurate predictions of complex compressible turbulence phenomena, such as the generation of jet noise that motivates the present effort. However, limited understanding of underlying physical mechanisms restricts the utility of such predictions since they do not, by themselves, indicate a route to design improvements. Gradient-based optimization using adjoints can circumvent the flow complexity to guide designs, though this is predicated on the availability of a sufficiently accurate solution of the forward and adjoint systems. These are challenging to obtain, since both the chaotic character of the turbulence and the typical use of discretizations near their resolution limits in order to efficiently represent its smaller scales will amplify any approximation errors made in the adjoint formulation. Formulating a practical exact adjoint that avoids such errors is especially challenging if it is to be compatible with state-of-the-art simulation methods used for the turbulent flow itself. Automatic differentiation (AD) can provide code to calculate a nominally exact adjoint, but existing general-purpose AD codes are inefficient to the point of being prohibitive for large-scale turbulence simulations. Here, we analyze the compressible flow equations as discretized using the same high-order workhorse methods used for many high-fidelity compressible turbulence simulations, and formulate a practical space-time discrete-adjoint method without changing the basic discretization. A key step is the definition of a particular discrete analog of the continuous norm that defines our cost functional; our selection leads directly to an efficient Runge-Kutta-like scheme, though it would be just first-order accurate if used outside the adjoint formulation for time integration, with finite-difference spatial operators for the adjoint system. Its computational cost only modestly exceeds that of the flow equations. We confirm that its
A practical discrete-adjoint method for high-fidelity compressible turbulence simulations
Vishnampet, Ramanathan; Bodony, Daniel J.; Freund, Jonathan B.
2015-03-15
Methods and computing hardware advances have enabled accurate predictions of complex compressible turbulence phenomena, such as the generation of jet noise that motivates the present effort. However, limited understanding of underlying physical mechanisms restricts the utility of such predictions since they do not, by themselves, indicate a route to design improvements. Gradient-based optimization using adjoints can circumvent the flow complexity to guide designs, though this is predicated on the availability of a sufficiently accurate solution of the forward and adjoint systems. These are challenging to obtain, since both the chaotic character of the turbulence and the typical use of discretizations near their resolution limits in order to efficiently represent its smaller scales will amplify any approximation errors made in the adjoint formulation. Formulating a practical exact adjoint that avoids such errors is especially challenging if it is to be compatible with state-of-the-art simulation methods used for the turbulent flow itself. Automatic differentiation (AD) can provide code to calculate a nominally exact adjoint, but existing general-purpose AD codes are inefficient to the point of being prohibitive for large-scale turbulence simulations. Here, we analyze the compressible flow equations as discretized using the same high-order workhorse methods used for many high-fidelity compressible turbulence simulations, and formulate a practical space–time discrete-adjoint method without changing the basic discretization. A key step is the definition of a particular discrete analog of the continuous norm that defines our cost functional; our selection leads directly to an efficient Runge–Kutta-like scheme, though it would be just first-order accurate if used outside the adjoint formulation for time integration, with finite-difference spatial operators for the adjoint system. Its computational cost only modestly exceeds that of the flow equations. We confirm that
An efficient parallel algorithm for the solution of a tridiagonal linear system of equations
NASA Technical Reports Server (NTRS)
Stone, H. S.
1971-01-01
Tridiagonal linear systems of equations are solved on conventional serial machines in a time proportional to N, where N is the number of equations. The conventional algorithms do not lend themselves directly to parallel computations on computers of the ILLIAC IV class, in the sense that they appear to be inherently serial. An efficient parallel algorithm is presented in which computation time grows as log sub 2 N. The algorithm is based on recursive doubling solutions of linear recurrence relations, and can be used to solve recurrence relations of all orders.
An efficient parallel algorithm for the solution of a tridiagonal linear system of equations.
NASA Technical Reports Server (NTRS)
Stone, H. S.
1973-01-01
Tridiagonal linear systems of equations can be solved on conventional serial machines in a time proportional to N, where N is the number of equations. The conventional algorithms do not lend themselves directly to parallel computation on computers of the Illiac IV class, in the sense that they appear to be inherently serial. An efficient parallel algorithm is presented in which computation time grows as log(sub-2) N. The algorithm is based on recursive doubling solutions of linear recurrence relations, and can be used to solve recurrence relations of all orders.
Adjoint tomography of the Middle East
NASA Astrophysics Data System (ADS)
Peter, D. B.; Savage, B.; Rodgers, A. J.; Tromp, J.
2010-12-01
Improvements in nuclear explosion monitoring require refined seismic models of the target region. In our study, we focus on the Middle East, spanning a region from Turkey to the west and West India to the east. This area represents a complex geologic and tectonic setting with sparse seismic data coverage. This has lead to diverging interpretations of crustal and underlying upper-mantle structure by different research groups, complicating seismic monitoring of the Middle East at regional distances. We evaluated an initial 3D seismic model of this region by computing full waveforms for several regional earthquakes by a spectral-element method. We measure traveltime and multitaper phase shifts between observed broadband data and synthetic seismograms for distinct seismic phases within selected time windows using a recently developed automated measurement algorithm. Based on the remaining misfits, we setup an iterative inversion procedure for a fully numerical 3D seismic tomography approach. In order to improve the initial 3D seismic model, the sensitivity to seismic structure of the traveltime and multitaper phase measurements for all available seismic network recordings is computed. As this represents a computationally very intensive task, we take advantage of a fully numerical adjoint approach by using the efficient software package SPECFEM3D_GLOBE on a dedicated cluster. We show examples of such sensitivity kernels for different seismic events and use them in a steepest descent approach to update the 3D seismic model, starting at longer periods between 60 s and up to 200 s and moving towards shorter periods of 11 s. We highlight various improvements in the initial seismic structure during the iterations in order to better fit regional seismic waveforms in the Middle East.
Adjoint tomography of the Middle East
NASA Astrophysics Data System (ADS)
Peter, D. B.; Savage, B.; Rodgers, A.; Morency, C.; Tromp, J.
2011-12-01
Improvements in nuclear explosion monitoring require refined seismic models of the target region. In our study, we focus on the Middle East, spanning a region from Turkey to the west and West India to the east. This area represents a complex geologic and tectonic setting with sparse seismic data coverage. This has lead to diverging interpretations of crustal and underlying upper-mantle structure by different research groups, complicating seismic monitoring of the Middle East at regional distances. We evaluated an initial 3D seismic model of this region by computing full waveforms for several regional earthquakes based on a spectral-element method. We measure traveltime and multitaper phase differences between observed broadband data and synthetic seismograms for distinct seismic phases within selected time windows using a recently developed automated measurement algorithm. Based on the remaining misfits, we setup an iterative inversion procedure for a fully numerical 3D seismic tomography approach. In order to improve the initial 3D seismic model, sensitivity to seismic structures of traveltime and multitaper phase measurements for all available seismic network recordings is computed. As this represents a computationally very intensive task, we take advantage of a fully numerical adjoint approach by using the efficient software package SPECFEM3D_GLOBE on a dedicated cluster. We show examples of such sensitivity kernels for different seismic events. All these `event kernels' are then summed, smoothed and further used in a preconditioned conjugate-gradient approach. Thus we iteratively update the 3D seismic model, starting at longer periods between 60~s and up to 150~s and moving towards shorter periods of 11~s. We highlight various improvements in the initial seismic structure during the iterations in order to better fit regional seismic waveforms in the Middle East.
Double-difference adjoint seismic tomography
NASA Astrophysics Data System (ADS)
Yuan, Yanhua O.; Simons, Frederik J.; Tromp, Jeroen
2016-06-01
We introduce a `double-difference' method for the inversion for seismic wavespeed structure based on adjoint tomography. Differences between seismic observations and model predictions at individual stations may arise from factors other than structural heterogeneity, such as errors in the assumed source-time function, inaccurate timings, and systematic uncertainties. To alleviate the corresponding nonuniqueness in the inverse problem, we construct differential measurements between stations, thereby reducing the influence of the source signature and systematic errors. We minimize the discrepancy between observations and simulations in terms of the differential measurements made on station pairs. We show how to implement the double-difference concept in adjoint tomography, both theoretically and in practice. We compare the sensitivities of absolute and differential measurements. The former provide absolute information on structure along the ray paths between stations and sources, whereas the latter explain relative (and thus higher-resolution) structural variations in areas close to the stations. Whereas in conventional tomography a measurement made on a single earthquake-station pair provides very limited structural information, in double-difference tomography one earthquake can actually resolve significant details of the structure. The double-difference methodology can be incorporated into the usual adjoint tomography workflow by simply pairing up all conventional measurements; the computational cost of the necessary adjoint simulations is largely unaffected. Rather than adding to the computational burden, the inversion of double-difference measurements merely modifies the construction of the adjoint sources for data assimilation.
NASA Astrophysics Data System (ADS)
Ersoy, Ozlem; Dag, Idris
2015-12-01
The solutions of the reaction-diffusion system are given by method of collocation based on the exponential B-splines. Thus the reaction-diffusion systemturns into an iterative banded algebraic matrix equation. Solution of the matrix equation is carried out byway of Thomas algorithm. The present methods test on both linear and nonlinear problems. The results are documented to compare with some earlier studies by use of L∞ and relative error norm for problems respectively.
Adjoint-based optimization for understanding and suppressing jet noise
NASA Astrophysics Data System (ADS)
Freund, Jonathan B.
2011-08-01
Advanced simulation tools, particularly large-eddy simulation techniques, are becoming capable of making quality predictions of jet noise for realistic nozzle geometries and at engineering relevant flow conditions. Increasing computer resources will be a key factor in improving these predictions still further. Quality prediction, however, is only a necessary condition for the use of such simulations in design optimization. Predictions do not themselves lead to quieter designs. They must be interpreted or harnessed in some way that leads to design improvements. As yet, such simulations have not yielded any simplifying principals that offer general design guidance. The turbulence mechanisms leading to jet noise remain poorly described in their complexity. In this light, we have implemented and demonstrated an aeroacoustic adjoint-based optimization technique that automatically calculates gradients that point the direction in which to adjust controls in order to improve designs. This is done with only a single flow solutions and a solution of an adjoint system, which is solved at computational cost comparable to that for the flow. Optimization requires iterations, but having the gradient information provided via the adjoint accelerates convergence in a manner that is insensitive to the number of parameters to be optimized. This paper, which follows from a presentation at the 2010 IUTAM Symposium on Computational Aero-Acoustics for Aircraft Noise Prediction, reviews recent and ongoing efforts by the author and co-workers. It provides a new formulation of the basic approach and demonstrates the approach on a series of model flows, culminating with a preliminary result for a turbulent jet.
Adjoint Techniques for Topology Optimization of Structures Under Damage Conditions
NASA Technical Reports Server (NTRS)
Akgun, Mehmet A.; Haftka, Raphael T.
2000-01-01
The objective of this cooperative agreement was to seek computationally efficient ways to optimize aerospace structures subject to damage tolerance criteria. Optimization was to involve sizing as well as topology optimization. The work was done in collaboration with Steve Scotti, Chauncey Wu and Joanne Walsh at the NASA Langley Research Center. Computation of constraint sensitivity is normally the most time-consuming step of an optimization procedure. The cooperative work first focused on this issue and implemented the adjoint method of sensitivity computation (Haftka and Gurdal, 1992) in an optimization code (runstream) written in Engineering Analysis Language (EAL). The method was implemented both for bar and plate elements including buckling sensitivity for the latter. Lumping of constraints was investigated as a means to reduce the computational cost. Adjoint sensitivity computation was developed and implemented for lumped stress and buckling constraints. Cost of the direct method and the adjoint method was compared for various structures with and without lumping. The results were reported in two papers (Akgun et al., 1998a and 1999). It is desirable to optimize topology of an aerospace structure subject to a large number of damage scenarios so that a damage tolerant structure is obtained. Including damage scenarios in the design procedure is critical in order to avoid large mass penalties at later stages (Haftka et al., 1983). A common method for topology optimization is that of compliance minimization (Bendsoe, 1995) which has not been used for damage tolerant design. In the present work, topology optimization is treated as a conventional problem aiming to minimize the weight subject to stress constraints. Multiple damage configurations (scenarios) are considered. Each configuration has its own structural stiffness matrix and, normally, requires factoring of the matrix and solution of the system of equations. Damage that is expected to be tolerated is local
Comparison of Ensemble and Adjoint Approaches to Variational Optimization of Observational Arrays
NASA Astrophysics Data System (ADS)
Nechaev, D.; Panteleev, G.; Yaremchuk, M.
2015-12-01
Comprehensive monitoring of the circulation in the Chukchi Sea and Bering Strait is one of the key prerequisites of the successful long-term forecast of the Arctic Ocean state. Since the number of continuously maintained observational platforms is restricted by logistical and political constraints, the configuration of such an observing system should be guided by an objective strategy that optimizes the observing system coverage, design, and the expenses of monitoring. The presented study addresses optimization of system consisting of a limited number of observational platforms with respect to reduction of the uncertainties in monitoring the volume/freshwater/heat transports through a set of key sections in the Chukchi Sea and Bering Strait. Variational algorithms for optimization of observational arrays are verified in the test bed of the set of 4Dvar optimized summer-fall circulations in the Pacific sector of the Arctic Ocean. The results of an optimization approach based on low-dimensional ensemble of model solutions is compared against a more conventional algorithm involving application of the tangent linear and adjoint models. Special attention is paid to the computational efficiency and portability of the optimization procedure.
A general algorithm for the solution of Kepler's equation for elliptic orbits
NASA Technical Reports Server (NTRS)
Ng, E. W.
1979-01-01
An efficient algorithm is presented for the solution of Kepler's equation f(E)=E-M-e sin E=0, where e is the eccentricity, M the mean anomaly and E the eccentric anomaly. This algorithm is based on simple initial approximations that are cubics in M, and an iterative scheme that is a slight generalization of the Newton-Raphson method. Extensive testing of this algorithm has been performed on the UNIVAC 1108 computer. Solutions for 20,000 pairs of values of e and M show that for single precision, 42.0% of the cases require one iteration, 57.8% two and 0.2% three. For double precision one additional iteration is required.
A generalized front marching algorithm for the solution of the eikonal equation
NASA Astrophysics Data System (ADS)
Covello, Paul; Rodrigue, Garry
2003-07-01
A new front marching algorithm for solving the eikonal equation is presented. An important property of the algorithm is that it can be used on nodes that are located on highly distorted grids or on nodes that are randomly located. When the nodes are located on an orthogonal grid, the method is first-order accurate and is shown to be a generalization of the front marching algorithm in (Proc. Natl. Acad. Sci. 93 (4) (1996) 1591). The accuracy of the method is also shown to be dependent on the principle curvature of the wave front solution. Numerical results on a variety of node configurations as well as on shadow, nonconvex and nondifferentiable solutions are presented.
ADGEN: ADjoint GENerator for computer models
Worley, B.A.; Pin, F.G.; Horwedel, J.E.; Oblow, E.M.
1989-05-01
This paper presents the development of a FORTRAN compiler and an associated supporting software library called ADGEN. ADGEN reads FORTRAN models as input and produces and enhanced version of the input model. The enhanced version reproduces the original model calculations but also has the capability to calculate derivatives of model results of interest with respect to any and all of the model data and input parameters. The method for calculating the derivatives and sensitivities is the adjoint method. Partial derivatives are calculated analytically using computer calculus and saved as elements of an adjoint matrix on direct assess storage. The total derivatives are calculated by solving an appropriate adjoint equation. ADGEN is applied to a major computer model of interest to the Low-Level Waste Community, the PRESTO-II model. PRESTO-II sample problem results reveal that ADGEN correctly calculates derivatives of response of interest with respect to 300 parameters. The execution time to create the adjoint matrix is a factor of 45 times the execution time of the reference sample problem. Once this matrix is determined, the derivatives with respect to 3000 parameters are calculated in a factor of 6.8 that of the reference model for each response of interest. For a single 3000 for determining these derivatives by parameter perturbations. The automation of the implementation of the adjoint technique for calculating derivatives and sensitivities eliminates the costly and manpower-intensive task of direct hand-implementation by reprogramming and thus makes the powerful adjoint technique more amenable for use in sensitivity analysis of existing models. 20 refs., 1 fig., 5 tabs.
FAST TRACK COMMUNICATION Quasi self-adjoint nonlinear wave equations
NASA Astrophysics Data System (ADS)
Ibragimov, N. H.; Torrisi, M.; Tracinà, R.
2010-11-01
In this paper we generalize the classification of self-adjoint second-order linear partial differential equation to a family of nonlinear wave equations with two independent variables. We find a class of quasi self-adjoint nonlinear equations which includes the self-adjoint linear equations as a particular case. The property of a differential equation to be quasi self-adjoint is important, e.g. for constructing conservation laws associated with symmetries of the differential equation.
Adjoint-Based Uncertainty Quantification with MCNP
Seifried, Jeffrey E.
2011-09-01
This work serves to quantify the instantaneous uncertainties in neutron transport simulations born from nuclear data and statistical counting uncertainties. Perturbation and adjoint theories are used to derive implicit sensitivity expressions. These expressions are transformed into forms that are convenient for construction with MCNP6, creating the ability to perform adjoint-based uncertainty quantification with MCNP6. These new tools are exercised on the depleted-uranium hybrid LIFE blanket, quantifying its sensitivities and uncertainties to important figures of merit. Overall, these uncertainty estimates are small (< 2%). Having quantified the sensitivities and uncertainties, physical understanding of the system is gained and some confidence in the simulation is acquired.
Mokeddem, Diab; Khellaf, Abdelhafid
2009-01-01
Optimal design problem are widely known by their multiple performance measures that are often competing with each other. In this paper, an optimal multiproduct batch chemical plant design is presented. The design is firstly formulated as a multiobjective optimization problem, to be solved using the well suited non dominating sorting genetic algorithm (NSGA-II). The NSGA-II have capability to achieve fine tuning of variables in determining a set of non dominating solutions distributed along the Pareto front in a single run of the algorithm. The NSGA-II ability to identify a set of optimal solutions provides the decision-maker DM with a complete picture of the optimal solution space to gain better and appropriate choices. Then an outranking with PROMETHEE II helps the decision-maker to finalize the selection of a best compromise. The effectiveness of NSGA-II method with multiojective optimization problem is illustrated through two carefully referenced examples. PMID:19543537
A structured multi-block solution-adaptive mesh algorithm with mesh quality assessment
NASA Technical Reports Server (NTRS)
Ingram, Clint L.; Laflin, Kelly R.; Mcrae, D. Scott
1995-01-01
The dynamic solution adaptive grid algorithm, DSAGA3D, is extended to automatically adapt 2-D structured multi-block grids, including adaption of the block boundaries. The extension is general, requiring only input data concerning block structure, connectivity, and boundary conditions. Imbedded grid singular points are permitted, but must be prevented from moving in space. Solutions for workshop cases 1 and 2 are obtained on multi-block grids and illustrate both increased resolution of and alignment with the solution. A mesh quality assessment criteria is proposed to determine how well a given mesh resolves and aligns with the solution obtained upon it. The criteria is used to evaluate the grid quality for solutions of workshop case 6 obtained on both static and dynamically adapted grids. The results indicate that this criteria shows promise as a means of evaluating resolution.
A deterministic annealing algorithm for approximating a solution of the min-bisection problem.
Dang, Chuangyin; Ma, Wei; Liang, Jiye
2009-01-01
The min-bisection problem is an NP-hard combinatorial optimization problem. In this paper an equivalent linearly constrained continuous optimization problem is formulated and an algorithm is proposed for approximating its solution. The algorithm is derived from the introduction of a logarithmic-cosine barrier function, where the barrier parameter behaves as temperature in an annealing procedure and decreases from a sufficiently large positive number to zero. The algorithm searches for a better solution in a feasible descent direction, which has a desired property that lower and upper bounds are always satisfied automatically if the step length is a number between zero and one. We prove that the algorithm converges to at least a local minimum point of the problem if a local minimum point of the barrier problem is generated for a sequence of descending values of the barrier parameter with a limit of zero. Numerical results show that the algorithm is much more efficient than two of the best existing heuristic methods for the min-bisection problem, Kernighan-Lin method with multiple starting points (MSKL) and multilevel graph partitioning scheme (MLGP). PMID:18995985
The finite state projection algorithm for the solution of the chemical master equation.
Munsky, Brian; Khammash, Mustafa
2006-01-28
This article introduces the finite state projection (FSP) method for use in the stochastic analysis of chemically reacting systems. One can describe the chemical populations of such systems with probability density vectors that evolve according to a set of linear ordinary differential equations known as the chemical master equation (CME). Unlike Monte Carlo methods such as the stochastic simulation algorithm (SSA) or tau leaping, the FSP directly solves or approximates the solution of the CME. If the CME describes a system that has a finite number of distinct population vectors, the FSP method provides an exact analytical solution. When an infinite or extremely large number of population variations is possible, the state space can be truncated, and the FSP method provides a certificate of accuracy for how closely the truncated space approximation matches the true solution. The proposed FSP algorithm systematically increases the projection space in order to meet prespecified tolerance in the total probability density error. For any system in which a sufficiently accurate FSP exists, the FSP algorithm is shown to converge in a finite number of steps. The FSP is utilized to solve two examples taken from the field of systems biology, and comparisons are made between the FSP, the SSA, and tau leaping algorithms. In both examples, the FSP outperforms the SSA in terms of accuracy as well as computational efficiency. Furthermore, due to very small molecular counts in these particular examples, the FSP also performs far more effectively than tau leaping methods. PMID:16460146
Adjoint optimization of natural convection problems: differentially heated cavity
NASA Astrophysics Data System (ADS)
Saglietti, Clio; Schlatter, Philipp; Monokrousos, Antonios; Henningson, Dan S.
2016-06-01
Optimization of natural convection-driven flows may provide significant improvements to the performance of cooling devices, but a theoretical investigation of such flows has been rarely done. The present paper illustrates an efficient gradient-based optimization method for analyzing such systems. We consider numerically the natural convection-driven flow in a differentially heated cavity with three Prandtl numbers (Pr=0.15{-}7 ) at super-critical conditions. All results and implementations were done with the spectral element code Nek5000. The flow is analyzed using linear direct and adjoint computations about a nonlinear base flow, extracting in particular optimal initial conditions using power iteration and the solution of the full adjoint direct eigenproblem. The cost function for both temperature and velocity is based on the kinetic energy and the concept of entransy, which yields a quadratic functional. Results are presented as a function of Prandtl number, time horizons and weights between kinetic energy and entransy. In particular, it is shown that the maximum transient growth is achieved at time horizons on the order of 5 time units for all cases, whereas for larger time horizons the adjoint mode is recovered as optimal initial condition. For smaller time horizons, the influence of the weights leads either to a concentric temperature distribution or to an initial condition pattern that opposes the mean shear and grows according to the Orr mechanism. For specific cases, it could also been shown that the computation of optimal initial conditions leads to a degenerate problem, with a potential loss of symmetry. In these situations, it turns out that any initial condition lying in a specific span of the eigenfunctions will yield exactly the same transient amplification. As a consequence, the power iteration converges very slowly and fails to extract all possible optimal initial conditions. According to the authors' knowledge, this behavior is illustrated here
An algorithm for constructing polynomial systems whose solution space characterizes quantum circuits
NASA Astrophysics Data System (ADS)
Gerdt, Vladimir P.; Severyanov, Vasily M.
2006-05-01
An algorithm and its first implementation in C# are presented for assembling arbitrary quantum circuits on the base of Hadamard and Toffoli gates and for constructing multivariate polynomial systems over the finite field Z II arising when applying the Feynman's sum-over-paths approach to quantum circuits. The matrix elements determined by a circuit can be computed by counting the number of common roots in Z II for the polynomial system associated with the circuit. To determine the number of solutions in Z II for the output polynomial system, one can use the Grobner bases method and the relevant algorithms for computing Grobner bases.
Solution algorithms for non-linear singularly perturbed optimal control problems
NASA Technical Reports Server (NTRS)
Ardema, M. D.
1983-01-01
The applicability and usefulness of several classical and other methods for solving the two-point boundary-value problem which arises in non-linear singularly perturbed optimal control are assessed. Specific algorithms of the Picard, Newton and averaging types are formally developed for this class of problem. The computational requirements associated with each algorithm are analysed and compared with the computational requirement of the method of matched asymptotic expansions. Approximate solutions to a linear and a non-linear problem are obtained by each method and compared.
NASA Astrophysics Data System (ADS)
Alfonso, Lester; Zamora, Jose; Cruz, Pedro
2015-04-01
The stochastic approach to coagulation considers the coalescence process going in a system of a finite number of particles enclosed in a finite volume. Within this approach, the full description of the system can be obtained from the solution of the multivariate master equation, which models the evolution of the probability distribution of the state vector for the number of particles of a given mass. Unfortunately, due to its complexity, only limited results were obtained for certain type of kernels and monodisperse initial conditions. In this work, a novel numerical algorithm for the solution of the multivariate master equation for stochastic coalescence that works for any type of kernels and initial conditions is introduced. The performance of the method was checked by comparing the numerically calculated particle mass spectrum with analytical solutions obtained for the constant and sum kernels, with an excellent correspondence between the analytical and numerical solutions. In order to increase the speedup of the algorithm, software parallelization techniques with OpenMP standard were used, along with an implementation in order to take advantage of new accelerator technologies. Simulations results show an important speedup of the parallelized algorithms. This study was funded by a grant from Consejo Nacional de Ciencia y Tecnologia de Mexico SEP-CONACYT CB-131879. The authors also thanks LUFAC® Computacion SA de CV for CPU time and all the support provided.
NASA Technical Reports Server (NTRS)
Ustinov, Eugene A.; Sunseri, Richard F.
2005-01-01
An approach is presented to the inversion of gravity fields based on evaluation of partials of observables with respect to gravity harmonics using the solution of adjoint problem of orbital dynamics of the spacecraft. Corresponding adjoint operator is derived directly from the linear operator of the linearized forward problem of orbital dynamics. The resulting adjoint problem is similar to the forward problem and can be solved by the same methods. For given highest degree N of gravity harmonics desired, this method involves integration of N adjoint solutions as compared to integration of N2 partials of the forward solution with respect to gravity harmonics in the conventional approach. Thus, for higher resolution gravity models, this approach becomes increasingly more effective in terms of computer resources as compared to the approach based on the solution of the forward problem of orbital dynamics.
NASA Technical Reports Server (NTRS)
Daso, E. O.
1986-01-01
An implicit approximate factorization algorithm is employed to quantify the parametric effects of Courant number and artificial smoothing on numerical solutions of the unsteady 3-D Euler equations for a windmilling propeller (low speed) flow field. The results show that propeller global or performance chracteristics vary strongly with Courant number and artificial dissipation parameters, though the variation is such less severe at high Courant numbers. Candidate sets of Courant number and dissipation parameters could result in parameter-dependent solutions. Parameter-independent numerical solutions can be obtained if low values of the dissipation parameter-time step ratio are used in the computations. Furthermore, it is realized that too much artificial damping can degrade numerical stability. Finally, it is demonstrated that highly resolved meshes may, in some cases, delay convergence, thereby suggesting some optimum cell size for a given flow solution. It is suspected that improper boundary treatment may account for the cell size constraint.
Investigation of ALEGRA shock hydrocode algorithms using an exact free surface jet flow solution.
Hanks, Bradley Wright.; Robinson, Allen Conrad
2014-01-01
Computational testing of the arbitrary Lagrangian-Eulerian shock physics code, ALEGRA, is presented using an exact solution that is very similar to a shaped charge jet flow. The solution is a steady, isentropic, subsonic free surface flow with significant compression and release and is provided as a steady state initial condition. There should be no shocks and no entropy production throughout the problem. The purpose of this test problem is to present a detailed and challenging computation in order to provide evidence for algorithmic strengths and weaknesses in ALEGRA which should be examined further. The results of this work are intended to be used to guide future algorithmic improvements in the spirit of test-driven development processes.
NASA Technical Reports Server (NTRS)
Nacozy, P. E.
1984-01-01
The equations of motion are developed for a perfectly flexible, inelastic tether with a satellite at its extremity. The tether is attached to a space vehicle in orbit. The tether is allowed to possess electrical conductivity. A numerical solution algorithm to provide the motion of the tether and satellite system is presented. The resulting differential equations can be solved by various existing standard numerical integration computer programs. The resulting differential equations allow the introduction of approximations that can lead to analytical, approximate general solutions. The differential equations allow more dynamical insight of the motion.
NASA Astrophysics Data System (ADS)
Lindegren, L.; Lammers, U.; Hobbs, D.; O'Mullane, W.; Bastian, U.; Hernández, J.
2012-02-01
Context. The Gaia satellite will observe about one billion stars and other point-like sources. The astrometric core solution will determine the astrometric parameters (position, parallax, and proper motion) for a subset of these sources, using a global solution approach which must also include a large number of parameters for the satellite attitude and optical instrument. The accurate and efficient implementation of this solution is an extremely demanding task, but crucial for the outcome of the mission. Aims: We aim to provide a comprehensive overview of the mathematical and physical models applicable to this solution, as well as its numerical and algorithmic framework. Methods: The astrometric core solution is a simultaneous least-squares estimation of about half a billion parameters, including the astrometric parameters for some 100 million well-behaved so-called primary sources. The global nature of the solution requires an iterative approach, which can be broken down into a small number of distinct processing blocks (source, attitude, calibration and global updating) and auxiliary processes (including the frame rotator and selection of primary sources). We describe each of these processes in some detail, formulate the underlying models, from which the observation equations are derived, and outline the adopted numerical solution methods with due consideration of robustness and the structure of the resulting system of equations. Appendices provide brief introductions to some important mathematical tools (quaternions and B-splines for the attitude representation, and a modified Cholesky algorithm for positive semidefinite problems) and discuss some complications expected in the real mission data. Results: A complete software system called AGIS (Astrometric Global Iterative Solution) is being built according to the methods described in the paper. Based on simulated data for 2 million primary sources we present some initial results, demonstrating the basic
Crane, N K; Parsons, I D; Hjelmstad, K D
2002-03-21
Adaptive mesh refinement selectively subdivides the elements of a coarse user supplied mesh to produce a fine mesh with reduced discretization error. Effective use of adaptive mesh refinement coupled with an a posteriori error estimator can produce a mesh that solves a problem to a given discretization error using far fewer elements than uniform refinement. A geometric multigrid solver uses increasingly finer discretizations of the same geometry to produce a very fast and numerically scalable solution to a set of linear equations. Adaptive mesh refinement is a natural method for creating the different meshes required by the multigrid solver. This paper describes the implementation of a scalable adaptive multigrid method on a distributed memory parallel computer. Results are presented that demonstrate the parallel performance of the methodology by solving a linear elastic rocket fuel deformation problem on an SGI Origin 3000. Two challenges must be met when implementing adaptive multigrid algorithms on massively parallel computing platforms. First, although the fine mesh for which the solution is desired may be large and scaled to the number of processors, the multigrid algorithm must also operate on much smaller fixed-size data sets on the coarse levels. Second, the mesh must be repartitioned as it is adapted to maintain good load balancing. In an adaptive multigrid algorithm, separate mesh levels may require separate partitioning, further complicating the load balance problem. This paper shows that, when the proper optimizations are made, parallel adaptive multigrid algorithms perform well on machines with several hundreds of processors.
A join algorithm for combining AND parallel solutions in AND/OR parallel systems
Ramkumar, B. ); Kale, L.V. )
1992-02-01
When two or more literals in the body of a Prolog clause are solved in (AND) parallel, their solutions need to be joined to compute solutions for the clause. This is often a difficult problem in parallel Prolog systems that exploit OR and independent AND parallelism in Prolog programs. In several AND/OR parallel systems proposed recently, this problem is side-stepped at the cost of unexploited OR parallelism in the program, in part due to the complexity of the backtracking algorithm beneath AND parallel branches. In some cases, the data dependency graphs used by these systems cannot represent all the exploitable independent AND parallelism known at compile time. In this paper, we describe the compile time analysis for an optimized join algorithm for supporting independent AND parallelism in logic programs efficiently without leaving and OR parallelism unexploited. We then discuss how this analysis can be used to yield very efficient runtime behavior. We also discuss problems associated with a tree representation of the search space when arbitrarily complex data dependency graphs are permitted. We describe how these problems can be resolved by mapping the search space onto data dependency graphs themselves. The algorithm has been implemented in a compiler for parallel Prolog based on the reduce-OR process model. The algorithm is suitable for the implementation of AND/OR systems on both shared and nonshared memory machines. Performance on benchmark programs.
Implementation of a block Lanczos algorithm for Eigenproblem solution of gyroscopic systems
NASA Technical Reports Server (NTRS)
Gupta, Kajal K.; Lawson, Charles L.
1987-01-01
The details of implementation of a general numerical procedure developed for the accurate and economical computation of natural frequencies and associated modes of any elastic structure rotating along an arbitrary axis are described. A block version of the Lanczos algorithm is derived for the solution that fully exploits associated matrix sparsity and employs only real numbers in all relevant computations. It is also capable of determining multiple roots and proves to be most efficient when compared to other, similar, exisiting techniques.
Two new exact solutions for relativistic perfect fluid spheres through Lake's algorithm
NASA Astrophysics Data System (ADS)
Maurya, S. K.; Gupta, Y. K.; Jasim, M. K.
2015-02-01
Two new exact solutions of Einstein's field equations for perfect fluid distribution are obtained using Lake's (Phys. Rev. D 67:104015, 2003) algorithm. The same are utilized to construct stellar models of physical relevance and possessing the maximum mass 2.6956 M Θ (quark star) and 0.9643 M Θ (white dwarfs) with the corresponding radius 20.5489 km and 3.1699 km respectively.
Coupling of Monte Carlo adjoint leakages with three-dimensional discrete ordinates forward fluences
Slater, C.O.; Lillie, R.A.; Johnson, J.O.; Simpson, D.B.
1998-04-01
A computer code, DRC3, has been developed for coupling Monte Carlo adjoint leakages with three-dimensional discrete ordinates forward fluences in order to solve a special category of geometrically-complex deep penetration shielding problems. The code extends the capabilities of earlier methods that coupled Monte Carlo adjoint leakages with two-dimensional discrete ordinates forward fluences. The problems involve the calculation of fluences and responses in a perturbation to an otherwise simple two- or three-dimensional radiation field. In general, the perturbation complicates the geometry such that it cannot be modeled exactly using any of the discrete ordinates geometry options and thus a direct discrete ordinates solution is not possible. Also, the calculation of radiation transport from the source to the perturbation involves deep penetration. One approach to solving such problems is to perform the calculations in three steps: (1) a forward discrete ordinates calculation, (2) a localized adjoint Monte Carlo calculation, and (3) a coupling of forward fluences from the first calculation with adjoint leakages from the second calculation to obtain the response of interest (fluence, dose, etc.). A description of this approach is presented along with results from test problems used to verify the method. The test problems that were selected could also be solved directly by the discrete ordinates method. The good agreement between the DRC3 results and the direct-solution results verify the correctness of DRC3.
Development of CO2 inversion system based on the adjoint of the global coupled transport model
NASA Astrophysics Data System (ADS)
Belikov, Dmitry; Maksyutov, Shamil; Chevallier, Frederic; Kaminski, Thomas; Ganshin, Alexander; Blessing, Simon
2014-05-01
We present the development of an inverse modeling system employing an adjoint of the global coupled transport model consisting of the National Institute for Environmental Studies (NIES) Eulerian transport model (TM) and the Lagrangian plume diffusion model (LPDM) FLEXPART. NIES TM is a three-dimensional atmospheric transport model, which solves the continuity equation for a number of atmospheric tracers on a grid spanning the entire globe. Spatial discretization is based on a reduced latitude-longitude grid and a hybrid sigma-isentropic coordinate in the vertical. NIES TM uses a horizontal resolution of 2.5°×2.5°. However, to resolve synoptic-scale tracer distributions and to have the ability to optimize fluxes at resolutions of 0.5° and higher we coupled NIES TM with the Lagrangian model FLEXPART. The Lagrangian component of the forward and adjoint models uses precalculated responses of the observed concentration to the surface fluxes and 3-D concentrations field simulated with the FLEXPART model. NIES TM and FLEXPART are driven by JRA-25/JCDAS reanalysis dataset. Construction of the adjoint of the Lagrangian part is less complicated, as LPDMs calculate the sensitivity of measurements to the surrounding emissions field by tracking a large number of "particles" backwards in time. Developing of the adjoint to Eulerian part was performed with automatic differentiation tool the Transformation of Algorithms in Fortran (TAF) software (http://www.FastOpt.com). This method leads to the discrete adjoint of NIES TM. The main advantage of the discrete adjoint is that the resulting gradients of the numerical cost function are exact, even for nonlinear algorithms. The overall advantages of our method are that: 1. No code modification of Lagrangian model is required, making it applicable to combination of global NIES TM and any Lagrangian model; 2. Once run, the Lagrangian output can be applied to any chemically neutral gas; 3. High-resolution results can be obtained over
NASA Astrophysics Data System (ADS)
Edwards, S.; Reuther, J.; Chattot, J. J.
The objective of this paper is to present a control theory approach for the design of airfoils in the presence of viscous compressible flows. A coupled system of the integral boundary layer and the Euler equations is solved to provide rapid flow simulations. An adjoint approach consistent with the complete coupled state equations is employed to obtain the sensitivities needed to drive a numerical optimization algorithm. Design to a target pressure distribution is demonstrated on an RAE 2822 airfoil at transonic speeds.
Application of Harmony Search algorithm to the solution of groundwater management models
NASA Astrophysics Data System (ADS)
Tamer Ayvaz, M.
2009-06-01
This study proposes a groundwater resources management model in which the solution is performed through a combined simulation-optimization model. A modular three-dimensional finite difference groundwater flow model, MODFLOW is used as the simulation model. This model is then combined with a Harmony Search (HS) optimization algorithm which is based on the musical process of searching for a perfect state of harmony. The performance of the proposed HS based management model is tested on three separate groundwater management problems: (i) maximization of total pumping from an aquifer (steady-state); (ii) minimization of the total pumping cost to satisfy the given demand (steady-state); and (iii) minimization of the pumping cost to satisfy the given demand for multiple management periods (transient). The sensitivity of HS algorithm is evaluated by performing a sensitivity analysis which aims to determine the impact of related solution parameters on convergence behavior. The results show that HS yields nearly same or better solutions than the previous solution methods and may be used to solve management problems in groundwater modeling.
NASA Technical Reports Server (NTRS)
Woods, Claudia M.; Brewe, David E.
1988-01-01
A numerical solution to a theoretical model of vapor cavitation in a dynamically loaded journal bearing is developed utilizing a multigrid iteration technique. The method is compared with a noniterative approach in terms of computational time and accuracy. The computational model is based on the Elrod algorithm, a control volume approach to the Reynolds equation which mimics the Jakobsson-Floberg and Olsson cavitation theory. Besides accounting for a moving cavitation boundary and conservation of mass at the boundary, it also conserves mass within the cavitated region via a smeared mass or striated flow extending to both surfaces in the film gap. The mixed nature of the equations (parabolic in the full film zone and hyperbolic in the cavitated zone) coupled with the dynamic aspects of the problem create interesting difficulties for the present solution approach. Emphasis is placed on the methods found to eliminate solution instabilities. Excellent results are obtained for both accuracy and reduction of computational time.
A new algorithm of ionospheric tomography——two-step solution
NASA Astrophysics Data System (ADS)
Wen, Debao
The inherent non-ideal geometry of ground-based global navigation satellite system (GNSS) observation stations distribution results in limited-angle tomographic inverse problems that are ill-posed. To cope with the above problem, a new tomographic algorithm, which is called two-step solution (TSS), is presented in this paper. In the new method, the electron density can be estimated by using two steps: 1) Phillips smoothing method (PSM) is first used to resolve the ill-posed problem in ionospheric tomography system; 2) The "coarse" solution of PSM is then input as the initial value of multiplicative algebraic reconstruction technique (MART) and improved by iterative mode. Numerical simulation experiment demonstrates that the two-step solution is feasible to GNSS-based ionospheric tomography and superior to PSM or MART alone.
NASA Technical Reports Server (NTRS)
Woods, C. M.; Brewe, D. E.
1989-01-01
A numerical solution to a theoretical model of vapor cavitation in a dynamically loaded journal bearing is developed utilizing a multigrid iteration technique. The method is compared with a noniterative approach in terms of computational time and accuracy. The computational model is based on the Elrod algorithm, a control volume approach to the Reynolds equation which mimics the Jakobsson-Floberg and Olsson cavitation theory. Besides accounting for a moving cavitation boundary and conservation of mass at the boundary, it also conserves mass within the cavitated region via a smeared mass or striated flow extending to both surfaces in the film gap. The mixed nature of the equations (parabolic in the full film zone and hyperbolic in the cavitated zone) coupled with the dynamic aspects of the problem create interesting difficulties for the present solution approach. Emphasis is placed on the methods found to eliminate solution instabilities. Excellent results are obtained for both accuracy and reduction of computational time.
Wing planform optimization via an adjoint method
NASA Astrophysics Data System (ADS)
Leoviriyakit, Kasidit
This dissertation focuses on the problem of wing planform optimization for transonic aircraft based on flow simulation using Computational Fluid Dynamics (CFD) combined with an adjoint-gradient based numerical optimization procedure. The adjoint method, traditionally used for wing section design has been extended to cover planform variations and to compute the sensitivities of the structural weight of both the wing section and planform variations. The two relevant disciplines accounted for are the aerodynamics and structural weight. A simplified structural weight model is used for the optimization. Results of a variety of long range transports indicate that significant improvement in both aerodynamics and structures can be achieved simultaneously. The proof-of-concept optimal results indicate large improvements for both drag and structural weight. The work is an "enabling step" towards a realistic automated wing designed by a computer.
McNunn, Gabriel S; Bryden, Kenneth M
2013-01-01
Tarjan's algorithm schedules the solution of systems of equations by noting the coupling and grouping between the equations. Simulating complex systems, e.g., advanced power plants, aerodynamic systems, or the multi-scale design of components, requires the linkage of large groups of coupled models. Currently, this is handled manually in systems modeling packages. That is, the analyst explicitly defines both the method and solution sequence necessary to couple the models. In small systems of models and equations this works well. However, as additional detail is needed across systems and across scales, the number of models grows rapidly. This precludes the manual assembly of large systems of federated models, particularly in systems composed of high fidelity models. This paper examines extending Tarjan's algorithm from sets of equations to sets of models. The proposed implementation of the algorithm is demonstrated using a small one-dimensional system of federated models representing the heat transfer and thermal stress in a gas turbine blade with thermal barrier coating. Enabling the rapid assembly and substitution of different models permits the rapid turnaround needed to support the “what-if” kinds of questions that arise in engineering design.
A conjugate gradient algorithm for the astrometric core solution of Gaia
NASA Astrophysics Data System (ADS)
Bombrun, A.; Lindegren, L.; Hobbs, D.; Holl, B.; Lammers, U.; Bastian, U.
2012-02-01
Context. The ESA space astrometry mission Gaia, planned to be launched in 2013, has been designed to make angular measurements on a global scale with micro-arcsecond accuracy. A key component of the data processing for Gaia is the astrometric core solution, which must implement an efficient and accurate numerical algorithm to solve the resulting, extremely large least-squares problem. The Astrometric Global Iterative Solution (AGIS) is a framework that allows to implement a range of different iterative solution schemes suitable for a scanning astrometric satellite. Aims: Our aim is to find a computationally efficient and numerically accurate iteration scheme for the astrometric solution, compatible with the AGIS framework, and a convergence criterion for deciding when to stop the iterations. Methods: We study an adaptation of the classical conjugate gradient (CG) algorithm, and compare it to the so-called simple iteration (SI) scheme that was previously known to converge for this problem, although very slowly. The different schemes are implemented within a software test bed for AGIS known as AGISLab. This allows to define, simulate and study scaled astrometric core solutions with a much smaller number of unknowns than in AGIS, and therefore to perform a large number of numerical experiments in a reasonable time. After successful testing in AGISLab, the CG scheme has been implemented also in AGIS. Results: The two algorithms CG and SI eventually converge to identical solutions, to within the numerical noise (of the order of 0.00001 micro-arcsec). These solutions are moreover independent of the starting values (initial star catalogue), and we conclude that they are equivalent to a rigorous least-squares estimation of the astrometric parameters. The CG scheme converges up to a factor four faster than SI in the tested cases, and in particular spatially correlated truncation errors are much more efficiently damped out with the CG scheme. While it appears to be difficult
NASA Astrophysics Data System (ADS)
Ren, L.; Liu, Q.
2012-12-01
We present multiple moment-tensor solution of the December 26, 2004 Sumatra earthquake based upon adjoint methods. An objective function Φ that measures the goodness of waveform fit between data and synthetics is minimized. Synthetics are calculated by spectral-element simulations (SPECFEM3D_GLOBE) in a 3D global earth model S362ANI to reduce the effect of heterogeneous structures. The Fréchet derivatives of Φ in the form δΦ = ∫T ∫VI(ɛ †ij)(x,T-t) δ(m_dot)ij(x,t)d3xdt, where δmij is the perturbation of moment density function and I(ɛ†ij)(x,T-t) denotes the time-integrated adjoint strain tensor, are calculated based upon adjoint methods implemented in SPECFEM3D_GLOBE. Our initial source model is obtained by monitoring the time-integrated adjoint strain tensors in the vicinity of the presumed source region. Source model parameters are iteratively updated by a preconditioned conjugate-gradient method to iteratively utilizing the calculated Φ and δΦ values. Our final inversion results show both similarities to and differences from previous source inversion results based on 1D background models.
NASA Astrophysics Data System (ADS)
Ren, L.; Liu, Q.; Hjörleifsdóttir, V.
2010-12-01
We present multiple moment-tensor solution of the Dec 26, 2004 Sumatra earthquake based upon the adjoint methods. An objective function Φ(m), where m is the multiple source model, measures the goodness of waveform fit between data and synthetics. The Fréchet derivatives of Φ in the form δΦ = ∫∫I(ɛ†)(x,T-t)δmij_dot(x,t)dVdt, where δmij is the source model perturbation and I(ɛ†)(x,T-t) denotes the time-integrated adjoint strain tensor, are calculated based upon adjoint methods and spectral-element simulations (SPECFEM3D_GLOBE) in a 3D global earth model S362ANI. Our initial source model is obtained independently by monitoring the time-integrated adjoint strain tensors around the presumed source region. We then utilize the Φ and δΦ calculations in a conjugate-gradient method to iteratively invert for the source model. Our final inversion results show both similarities with and differences to previous source inversion results based on 1D earth models.
Supersymmetric descendants of self-adjointly extended quantum mechanical Hamiltonians
NASA Astrophysics Data System (ADS)
Al-Hashimi, M. H.; Salman, M.; Shalaby, A.; Wiese, U.-J.
2013-10-01
We consider the descendants of self-adjointly extended Hamiltonians in supersymmetric quantum mechanics on a half-line, on an interval, and on a punctured line or interval. While there is a 4-parameter family of self-adjointly extended Hamiltonians on a punctured line, only a 3-parameter sub-family has supersymmetric descendants that are themselves self-adjoint. We also address the self-adjointness of an operator related to the supercharge, and point out that only a sub-class of its most general self-adjoint extensions is physical. Besides a general characterization of self-adjoint extensions and their supersymmetric descendants, we explicitly consider concrete examples, including a particle in a box with general boundary conditions, with and without an additional point interaction. We also discuss bulk-boundary resonances and their manifestation in the supersymmetric descendant.
Bu, Sunyoung; Huang, Jingfang; Boyer, Treavor H.; Miller, Cass T.
2010-01-01
The focus of this work is on the modeling of an ion exchange process that occurs in drinking water treatment applications. The model formulation consists of a two-scale model in which a set of microscale diffusion equations representing ion exchange resin particles that vary in size and age are coupled through a boundary condition with a macroscopic ordinary differential equation (ODE), which represents the concentration of a species in a well-mixed reactor. We introduce a new age-averaged model (AAM) that averages all ion exchange particle ages for a given size particle to avoid the expensive Monte-Carlo simulation associated with previous modeling applications. We discuss two different numerical schemes to approximate both the original Monte Carlo algorithm and the new AAM for this two-scale problem. The first scheme is based on the finite element formulation in space coupled with an existing backward-difference-formula-based ODE solver in time. The second scheme uses an integral equation based Krylov deferred correction (KDC) method and a fast elliptic solver (FES) for the resulting elliptic equations. Numerical results are presented to validate the new AAM algorithm, which is also shown to be more computationally efficient than the original Monte Carlo algorithm. We also demonstrate that the higher order KDC scheme is more efficient than the traditional finite element solution approach and this advantage becomes increasingly important as the desired accuracy of the solution increases. We also discuss issues of smoothness, which affect the efficiency of the KDC-FES approach, and outline additional algorithmic changes that would further improve the efficiency of these developing methods for a wide range of applications. PMID:20577570
NASA Astrophysics Data System (ADS)
Bu, Sunyoung; Huang, Jingfang; Boyer, Treavor H.; Miller, Cass T.
2010-07-01
The focus of this work is on the modeling of an ion exchange process that occurs in drinking water treatment applications. The model formulation consists of a two-scale model in which a set of microscale diffusion equations representing ion exchange resin particles that vary in size and age are coupled through a boundary condition with a macroscopic ordinary differential equation (ODE), which represents the concentration of a species in a well-mixed reactor. We introduce a new age-averaged model (AAM) that averages all ion exchange particle ages for a given size particle to avoid the expensive Monte-Carlo simulation associated with previous modeling applications. We discuss two different numerical schemes to approximate both the original Monte-Carlo algorithm and the new AAM for this two-scale problem. The first scheme is based on the finite element formulation in space coupled with an existing backward difference formula-based ODE solver in time. The second scheme uses an integral equation based Krylov deferred correction (KDC) method and a fast elliptic solver (FES) for the resulting elliptic equations. Numerical results are presented to validate the new AAM algorithm, which is also shown to be more computationally efficient than the original Monte-Carlo algorithm. We also demonstrate that the higher order KDC scheme is more efficient than the traditional finite element solution approach and this advantage becomes increasingly important as the desired accuracy of the solution increases. We also discuss issues of smoothness, which affect the efficiency of the KDC-FES approach, and outline additional algorithmic changes that would further improve the efficiency of these developing methods for a wide range of applications.
Cesari, G.
1994-12-31
The aim of this paper is to analyze experimentally the quality of the solution obtained with dissection algorithms applied to the geometric Traveling Salesman Problem. Starting from Karp`s results. We apply a divide and conquer strategy, first dividing the plane into subregions where we calculate optimal subtours and then merging these subtours to obtain the final tour. The analysis is restricted to problem instances where points are uniformly distributed in the unit square. For relatively small sets of cities we analyze the quality of the solution by calculating the length of the optimal tour and by comparing it with our approximate solution. When the problem instance is too large we perform an asymptotical analysis estimating the length of the optimal tour. We apply the same dissection strategy also to classical heuristics by calculating approximate subtours and by comparing the results with the average quality of the heuristic. Our main result is the estimate of the rate of convergence of the approximate solution to the optimal solution as a function of the number of dissection steps, of the criterion used for the plane division and of the quality of the subtours. We have implemented our programs on MUSIC (MUlti Signal processor system with Intelligent Communication), a Single-Program-Multiple-Data parallel computer with distributed memory developed at the ETH Zurich.
NASA Technical Reports Server (NTRS)
Gunzburger, M. D.; Nicolaides, R. A.
1986-01-01
Substructuring methods are in common use in mechanics problems where typically the associated linear systems of algebraic equations are positive definite. Here these methods are extended to problems which lead to nonpositive definite, nonsymmetric matrices. The extension is based on an algorithm which carries out the block Gauss elimination procedure without the need for interchanges even when a pivot matrix is singular. Examples are provided wherein the method is used in connection with finite element solutions of the stationary Stokes equations and the Helmholtz equation, and dual methods for second-order elliptic equations.
ERIC Educational Resources Information Center
Adachi, Kohei
2013-01-01
Rubin and Thayer ("Psychometrika," 47:69-76, 1982) proposed the EM algorithm for exploratory and confirmatory maximum likelihood factor analysis. In this paper, we prove the following fact: the EM algorithm always gives a proper solution with positive unique variances and factor correlations with absolute values that do not exceed one, when the…
NASA Technical Reports Server (NTRS)
Doxley, Charles A.
2016-01-01
In the current world of applications that use reconfigurable technology implemented on field programmable gate arrays (FPGAs), there is a need for flexible architectures that can grow as the systems evolve. A project has limited resources and a fixed set of requirements that development efforts are tasked to meet. Designers must develop robust solutions that practically meet the current customer demands and also have the ability to grow for future performance. This paper describes the development of a high speed serial data streaming algorithm that allows for transmission of multiple data channels over a single serial link. The technique has the ability to change to meet new applications developed for future design considerations. This approach uses the Xilinx Serial RapidIO LOGICORE Solution to implement a flexible infrastructure to meet the current project requirements with the ability to adapt future system designs.
Introduction of Parallel GPGPU Acceleration Algorithms for the Solution of Radiative Transfer
NASA Technical Reports Server (NTRS)
Godoy, William F.; Liu, Xu
2011-01-01
General-purpose computing on graphics processing units (GPGPU) is a recent technique that allows the parallel graphics processing unit (GPU) to accelerate calculations performed sequentially by the central processing unit (CPU). To introduce GPGPU to radiative transfer, the Gauss-Seidel solution of the well-known expressions for 1-D and 3-D homogeneous, isotropic media is selected as a test case. Different algorithms are introduced to balance memory and GPU-CPU communication, critical aspects of GPGPU. Results show that speed-ups of one to two orders of magnitude are obtained when compared to sequential solutions. The underlying value of GPGPU is its potential extension in radiative solvers (e.g., Monte Carlo, discrete ordinates) at a minimal learning curve.
Efficient solution of liquid state integral equations using the Newton-GMRES algorithm
NASA Astrophysics Data System (ADS)
Booth, Michael J.; Schlijper, A. G.; Scales, L. E.; Haymet, A. D. J.
1999-06-01
We present examples of the accurate, robust and efficient solution of Ornstein-Zernike type integral equations which describe the structure of both homogeneous and inhomogeneous fluids. In this work we use the Newton-GMRES algorithm as implemented in the public-domain nonlinear Krylov solvers NKSOL [ P. Brown, Y. Saad, SIAM J. Sci. Stat. Comput. 11 (1990) 450] and NITSOL [ M. Pernice, H.F. Walker, SIAM J. Sci. Comput. 19 (1998) 302]. We compare and contrast this method with more traditional approaches in the literature, using Picard iteration (successive-substitution) and hybrid Newton-Raphson and Picard methods, and a recent vector extrapolation method [ H.H.H. Homeier, S. Rast, H. Krienke, Comput. Phys. Commun. 92 (1995) 188]. We find that both the performance and ease of implementation of these nonlinear solvers recommend them for the solution of this class of problem.
A Solution Method of Scheduling Problem with Worker Allocation by a Genetic Algorithm
NASA Astrophysics Data System (ADS)
Osawa, Akira; Ida, Kenichi
In a scheduling problem with worker allocation (SPWA) proposed by Iima et al, the worker's skill level to each machine is all the same. However, each worker has a different skill level for each machine in the real world. For that reason, we propose a new model of SPWA in which a worker has the different skill level to each machine. To solve the problem, we propose a new GA for SPWA consisting of the following new three procedures, shortening of idle time, modifying infeasible solution to feasible solution, and a new selection method for GA. The effectiveness of the proposed algorithm is clarified by numerical experiments using benchmark problems for job-shop scheduling.
Algorithmic Construction of Exact Solutions for Neutral Static Perfect Fluid Spheres
NASA Astrophysics Data System (ADS)
Hansraj, Sudan; Krupanandan, Daniel
2013-07-01
Although it ranks amongst the oldest of problems in classical general relativity, the challenge of finding new exact solutions for spherically symmetric perfect fluid spacetimes is still ongoing because of a paucity of solutions which exhibit the necessary qualitative features compatible with observational evidence. The problem amounts to solving a system of three partial differential equations in four variables, which means that any one of four geometric or dynamical quantities must be specified at the outset and the others should follow by integration. The condition of pressure isotropy yields a differential equation that may be interpreted as second-order in one of the space variables or also as first-order Ricatti type in the other space variable. This second option has been fruitful in allowing us to construct an algorithm to generate a complete solution to the Einstein field equations once a geometric variable is specified ab initio. We then demonstrate the construction of previously unreported solutions and examine these for physical plausibility as candidates to represent real matter. In particular we demand positive definiteness of pressure, density as well as a subluminal sound speed. Additionally, we require the existence of a hypersurface of vanishing pressure to identify a radius for the closed distribution of fluid. Finally, we examine the energy conditions. We exhibit models which display all of these elementary physical requirements.
NASA Astrophysics Data System (ADS)
Marotzke, Jochem; Giering, Ralf; Zhang, Kate Q.; Stammer, Detlef; Hill, Chris; Lee, Tong
1999-12-01
We first describe the principles and practical considerations behind the computer generation of the adjoint to the Massachusetts Institute of Technology ocean general circulation model (GCM) using R. Giering's software tool Tangent-Linear and Adjoint Model Compiler (TAMC). The TAMC's recipe for (FORTRAN-) line-by-line generation of adjoint code is explained by interpreting an adjoint model strictly as the operator that gives the sensitivity of the output of a model to its input. Then, the sensitivity of 1993 annual mean heat transport across 29°N in the Atlantic, to the hydrography on January 1, 1993, is calculated from a global solution of the GCM. The "kinematic sensitivity" to initial temperature variations is isolated, showing how the latter would influence heat transport if they did not affect the density and hence the flow. Over 1 year the heat transport at 29°N is influenced kinematically from regions up to 20° upstream in the western boundary current and up to 5° upstream in the interior. In contrast, the dynamical influences of initial temperature (and salinity) perturbations spread from as far as the rim of the Labrador Sea to the 29°N section along the western boundary. The sensitivities calculated with the adjoint compare excellently to those from a perturbation calculation with the dynamical model. Perturbations in initial interior salinity influence meridional overturning and heat transport when they have propagated to the western boundary and can thus influence the integrated east-west density difference. Our results support the notion that boundary monitoring of meridional mass and heat transports is feasible.
Optimization of the K-means algorithm for the solution of high dimensional instances
NASA Astrophysics Data System (ADS)
Pérez, Joaquín; Pazos, Rodolfo; Olivares, Víctor; Hidalgo, Miguel; Ruiz, Jorge; Martínez, Alicia; Almanza, Nelva; González, Moisés
2016-06-01
This paper addresses the problem of clustering instances with a high number of dimensions. In particular, a new heuristic for reducing the complexity of the K-means algorithm is proposed. Traditionally, there are two approaches that deal with the clustering of instances with high dimensionality. The first executes a preprocessing step to remove those attributes of limited importance. The second, called divide and conquer, creates subsets that are clustered separately and later their results are integrated through post-processing. In contrast, this paper proposes a new solution which consists of the reduction of distance calculations from the objects to the centroids at the classification step. This heuristic is derived from the visual observation of the clustering process of K-means, in which it was found that the objects can only migrate to adjacent clusters without crossing distant clusters. Therefore, this heuristic can significantly reduce the number of distance calculations from an object to the centroids of the potential clusters that it may be classified to. To validate the proposed heuristic, it was designed a set of experiments with synthetic and high dimensional instances. One of the most notable results was obtained for an instance of 25,000 objects and 200 dimensions, where its execution time was reduced up to 96.5% and the quality of the solution decreased by only 0.24% when compared to the K-means algorithm.
A modern solver framework to manage solution algorithms in the Community Earth System Model
Evans, Katherine J; Worley, Patrick H; Nichols, Dr Jeff A; WhiteIII, James B; Salinger, Andy; Price, Stephen; Lemieux, Jean-Francois; Lipscomb, William; Perego, Mauro; Vertenstein, Mariana; Edwards, Jim
2012-01-01
Global Earth-system models (ESM) can now produce simulations that resolve ~50 km features and include finer-scale, interacting physical processes. In order to achieve these scale-length solutions, ESMs require smaller time steps, which limits parallel performance. Solution methods that overcome these bottlenecks can be quite intricate, and there is no single set of algorithms that perform well across the range of problems of interest. This creates significant implementation challenges, which is further compounded by complexity of ESMs. Therefore, prototyping and evaluating new algorithms in these models requires a software framework that is flexible, extensible, and easily introduced into the existing software. We describe our efforts to create a parallel solver framework that links the Trilinos library of solvers to Glimmer-CISM, a continental ice sheet model used in the Community Earth System Model (CESM). We demonstrate this framework within both current and developmental versions of Glimmer-CISM and provide strategies for its integration into the rest of the CESM.
Wang, H.; Man, S.; Ewing, R.E.; Qin, G.; Lyons, S.L.; Al-Lawatia, M.
1999-06-10
Many difficult problems arise in the numerical simulation of fluid flow processes within porous media in petroleum reservoir simulation and in subsurface contaminant transport and remediation. The authors develop a family of Eulerian-Lagrangian localized adjoint methods for the solution of the initial-boundary value problems for first-order advection-reaction equations on general multi-dimensional domains. Different tracking algorithms, including the Euler and Runge-Kutta algorithms, are used. The derived schemes, which are full mass conservative, naturally incorporate inflow boundary conditions into their formulations and do not need any artificial outflow boundary conditions. Moreover, they have regularly structured, well-conditioned, symmetric, and positive-definite coefficient matrices, which can be efficiently solved by the conjugate gradient method in an optimal order number of iterations without any preconditioning needed. Numerical results are presented to compare the performance of the ELLAM schemes with many well studied and widely used methods, including the upwind finite difference method, the Galerkin and the Petrov-Galerkin finite element methods with backward-Euler or Crank-Nicolson temporal discretization, the streamline diffusion finite element methods, the monotonic upstream-centered scheme for conservation laws (MUSCL), and the Minmod scheme.
NASA Technical Reports Server (NTRS)
Batina, John T.
1992-01-01
A time-accurate approximate-factorization (AF) algorithm is described for solution of the three-dimensional unsteady transonic small-disturbance equation. The AF algorithm consists of a time-linearization procedure coupled with a subiteration technique. The algorithm is the basis for the Computational Aeroelasticity Program-Transonic Small Disturbance (CAP-TSD) computer code, which was developed for the analysis of unsteady aerodynamics and aeroelasticity of realistic aircraft configurations. The paper describes details on the governing flow equations and boundary conditions, with an emphasis on documenting the finite-difference formulas of the AF algorithm.
Adjoint sensitivity analysis of an ultrawideband antenna
Stephanson, M B; White, D A
2011-07-28
The frequency domain finite element method using H(curl)-conforming finite elements is a robust technique for full-wave analysis of antennas. As computers become more powerful, it is becoming feasible to not only predict antenna performance, but also to compute sensitivity of antenna performance with respect to multiple parameters. This sensitivity information can then be used for optimization of the design or specification of manufacturing tolerances. In this paper we review the Adjoint Method for sensitivity calculation, and apply it to the problem of optimizing a Ultrawideband antenna.
Improved Adjoint-Operator Learning For A Neural Network
NASA Technical Reports Server (NTRS)
Toomarian, Nikzad; Barhen, Jacob
1995-01-01
Improved method of adjoint-operator learning reduces amount of computation and associated computational memory needed to make electronic neural network learn temporally varying pattern (e.g., to recognize moving object in image) in real time. Method extension of method described in "Adjoint-Operator Learning for a Neural Network" (NPO-18352).
Adjoint methods for external beam inverse treatment planning
NASA Astrophysics Data System (ADS)
Kowalok, Michael E.
Forward and adjoint radiation transport methods may both be used to determine the dosimetric relationship between source parameters and voxel elements of a phantom. Forward methods consider one specific tuple of source parameters and calculate the response in all voxels of interest. This response is often cast as the dose delivered per unit source-weight. Adjoint transport methods, conversely, consider one particular voxel and calculate the response of that voxel in relation to all possible source parameters. In this regard, adjoint methods provide an "adjoint function" in addition to a dose value. Although the dose is for a single voxel only, the adjoint function illustrates the source parameters, (e.g. beam positions and directions) that are most important to delivering the dose to that voxel. In this regard, adjoint methods of analysis lend themselves in a natural way to optimization problems and perturbation studies. This work investigates the utility of adjoint analytic methods for treatment planning and for Monte Carlo dose calculations. Various methods for implementing this approach are discussed, along with their strengths and weaknesses. The complementary nature of adjoint and forward techniques is illustrated and exploited. Also, several features of the Monte Carlo codes MCNP and MCNPX are reviewed for treatment planning applications.
Supersymmetric descendants of self-adjointly extended quantum mechanical Hamiltonians
Al-Hashimi, M.H.; Salman, M.; Shalaby, A.; Wiese, U.-J.
2013-10-15
We consider the descendants of self-adjointly extended Hamiltonians in supersymmetric quantum mechanics on a half-line, on an interval, and on a punctured line or interval. While there is a 4-parameter family of self-adjointly extended Hamiltonians on a punctured line, only a 3-parameter sub-family has supersymmetric descendants that are themselves self-adjoint. We also address the self-adjointness of an operator related to the supercharge, and point out that only a sub-class of its most general self-adjoint extensions is physical. Besides a general characterization of self-adjoint extensions and their supersymmetric descendants, we explicitly consider concrete examples, including a particle in a box with general boundary conditions, with and without an additional point interaction. We also discuss bulk-boundary resonances and their manifestation in the supersymmetric descendant. -- Highlights: •Self-adjoint extension theory and contact interactions. •Application of self-adjoint extensions to supersymmetry. •Contact interactions in finite volume with Robin boundary condition.
The compressible adjoint equations in geodynamics: equations and numerical assessment
NASA Astrophysics Data System (ADS)
Ghelichkhan, Siavash; Bunge, Hans-Peter
2016-04-01
The adjoint method is a powerful means to obtain gradient information in a mantle convection model relative to past flow structure. While the adjoint equations in geodynamics have been derived for the conservation equations of mantle flow in their incompressible form, the applicability of this approximation to Earth is limited, because density increases by almost a factor of two from the surface to the Core Mantle Boundary. Here we introduce the compressible adjoint equations for the conservation equations in the anelastic-liquid approximation. Our derivation applies an operator formulation in Hilbert spaces, to connect to recent work in seismology (Fichtner et al (2006)) and geodynamics (Horbach et al (2014)), where the approach was used to derive the adjoint equations for the wave equation and incompressible mantle flow. We present numerical tests of the newly derived equations based on twin experiments, focusing on three simulations. A first, termed Compressible, assumes the compressible forward and adjoint equations, and represents the consistent means of including compressibility effects. A second, termed Mixed, applies the compressible forward equation, but ignores compressibility effects in the adjoint equations, where the incompressible equations are used instead. A third simulation, termed Incompressible, neglects compressibility effects entirely in the forward and adjoint equations relative to the reference twin. The compressible and mixed formulations successfully restore earlier mantle flow structure, while the incompressible formulation yields noticeable artifacts. Our results suggest the use of a compressible formulation, when applying the adjoint method to seismically derived mantle heterogeneity structure.
Efficient checkpointing schemes for depletion perturbation solutions on memory-limited architectures
Stripling, H. F.; Adams, M. L.; Hawkins, W. D.
2013-07-01
We describe a methodology for decreasing the memory footprint and machine I/O load associated with the need to access a forward solution during an adjoint solve. Specifically, we are interested in the depletion perturbation equations, where terms in the adjoint Bateman and transport equations depend on the forward flux solution. Checkpointing is the procedure of storing snapshots of the forward solution to disk and using these snapshots to recompute the parts of the forward solution that are necessary for the adjoint solve. For large problems, however, the storage cost of just a few copies of an angular flux vector can exceed the available RAM on the host machine. We propose a methodology that does not checkpoint the angular flux vector; instead, we write and store converged source moments, which are typically of a much lower dimension than the angular flux solution. This reduces the memory footprint and I/O load of the problem, but requires that we perform single sweeps to reconstruct flux vectors on demand. We argue that this trade-off is exactly the kind of algorithm that will scale on advanced, memory-limited architectures. We analyze the cost, in terms of FLOPS and memory footprint, of five checkpointing schemes. We also provide computational results that support the analysis and show that the memory-for-work trade off does improve time to solution. (authors)
Modeling the Pulse Line Ion Accelerator (PLIA): an algorithm for quasi-static field solution.
Friedman, A; Briggs, R J; Grote, D P; Henestroza, E; Waldron, W L
2007-06-18
The Pulse-Line Ion Accelerator (PLIA) is a helical distributed transmission line. A rising pulse applied to the upstream end appears as a moving spatial voltage ramp, on which an ion pulse can be accelerated. This is a promising approach to acceleration and longitudinal compression of an ion beam at high line charge density. In most of the studies carried out to date, using both a simple code for longitudinal beam dynamics and the Warp PIC code, a circuit model for the wave behavior was employed; in Warp, the helix I and V are source terms in elliptic equations for E and B. However, it appears possible to obtain improved fidelity using a ''sheath helix'' model in the quasi-static limit. Here we describe an algorithmic approach that may be used to effect such a solution.
NASA Astrophysics Data System (ADS)
Kuraz, Michal
2016-06-01
This paper presents pseudo-deterministic catchment runoff model based on the Richards equation model [1] - the governing equation for the subsurface flow. The subsurface flow in a catchment is described here by two-dimensional variably saturated flow (unsaturated and saturated). The governing equation is the Richards equation with a slight modification of the time derivative term as considered e.g. by Neuman [2]. The nonlinear nature of this problem appears in unsaturated zone only, however the delineation of the saturated zone boundary is a nonlinear computationally expensive issue. The simple one-dimensional Boussinesq equation was used here as a rough estimator of the saturated zone boundary. With this estimate the dd-adaptivity algorithm (see Kuraz et al. [4, 5, 6]) could always start with an optimal subdomain split, so it is now possible to avoid solutions of huge systems of linear equations in the initial iteration level of our Richards equation based runoff model.
NASA Astrophysics Data System (ADS)
Barrett, Steven R. H.; Britter, Rex E.
Predicting long-term mean pollutant concentrations in the vicinity of airports, roads and other industrial sources are frequently of concern in regulatory and public health contexts. Many emissions are represented geometrically as ground-level line or area sources. Well developed modelling tools such as AERMOD and ADMS are able to model dispersion from finite (i.e. non-point) sources with considerable accuracy, drawing upon an up-to-date understanding of boundary layer behaviour. Due to mathematical difficulties associated with line and area sources, computationally expensive numerical integration schemes have been developed. For example, some models decompose area sources into a large number of line sources orthogonal to the mean wind direction, for which an analytical (Gaussian) solution exists. Models also employ a time-series approach, which involves computing mean pollutant concentrations for every hour over one or more years of meteorological data. This can give rise to computer runtimes of several days for assessment of a site. While this may be acceptable for assessment of a single industrial complex, airport, etc., this level of computational cost precludes national or international policy assessments at the level of detail available with dispersion modelling. In this paper, we extend previous work [S.R.H. Barrett, R.E. Britter, 2008. Development of algorithms and approximations for rapid operational air quality modelling. Atmospheric Environment 42 (2008) 8105-8111] to line and area sources. We introduce approximations which allow for the development of new analytical solutions for long-term mean dispersion from line and area sources, based on hypergeometric functions. We describe how these solutions can be parameterized from a single point source run from an existing advanced dispersion model, thereby accounting for all processes modelled in the more costly algorithms. The parameterization method combined with the analytical solutions for long-term mean
NASA Astrophysics Data System (ADS)
Belikov, Dmitry A.; Maksyutov, Shamil; Yaremchuk, Alexey; Ganshin, Alexander; Kaminski, Thomas; Blessing, Simon; Sasakawa, Motoki; Gomez-Pelaez, Angel J.; Starchenko, Alexander
2016-02-01
We present the development of the Adjoint of the Global Eulerian-Lagrangian Coupled Atmospheric (A-GELCA) model that consists of the National Institute for Environmental Studies (NIES) model as an Eulerian three-dimensional transport model (TM), and FLEXPART (FLEXible PARTicle dispersion model) as the Lagrangian Particle Dispersion Model (LPDM). The forward tangent linear and adjoint components of the Eulerian model were constructed directly from the original NIES TM code using an automatic differentiation tool known as TAF (Transformation of Algorithms in Fortran; http://www.FastOpt.com, with additional manual pre- and post-processing aimed at improving transparency and clarity of the code and optimizing the performance of the computing, including MPI (Message Passing Interface). The Lagrangian component did not require any code modification, as LPDMs are self-adjoint and track a significant number of particles backward in time in order to calculate the sensitivity of the observations to the neighboring emission areas. The constructed Eulerian adjoint was coupled with the Lagrangian component at a time boundary in the global domain. The simulations presented in this work were performed using the A-GELCA model in forward and adjoint modes. The forward simulation shows that the coupled model improves reproduction of the seasonal cycle and short-term variability of CO2. Mean bias and standard deviation for five of the six Siberian sites considered decrease roughly by 1 ppm when using the coupled model. The adjoint of the Eulerian model was shown, through several numerical tests, to be very accurate (within machine epsilon with mismatch around to ±6 e-14) compared to direct forward sensitivity calculations. The developed adjoint of the coupled model combines the flux conservation and stability of an Eulerian discrete adjoint formulation with the flexibility, accuracy, and high resolution of a Lagrangian backward trajectory formulation. A-GELCA will be incorporated
Adjoint equations and analysis of complex systems: Application to virus infection modelling
NASA Astrophysics Data System (ADS)
Marchuk, G. I.; Shutyaev, V.; Bocharov, G.
2005-12-01
Recent development of applied mathematics is characterized by ever increasing attempts to apply the modelling and computational approaches across various areas of the life sciences. The need for a rigorous analysis of the complex system dynamics in immunology has been recognized since more than three decades ago. The aim of the present paper is to draw attention to the method of adjoint equations. The methodology enables to obtain information about physical processes and examine the sensitivity of complex dynamical systems. This provides a basis for a better understanding of the causal relationships between the immune system's performance and its parameters and helps to improve the experimental design in the solution of applied problems. We show how the adjoint equations can be used to explain the changes in hepatitis B virus infection dynamics between individual patients.
Development of Gis Tool for the Solution of Minimum Spanning Tree Problem using Prim's Algorithm
NASA Astrophysics Data System (ADS)
Dutta, S.; Patra, D.; Shankar, H.; Alok Verma, P.
2014-11-01
minimum spanning tree (MST) of a connected, undirected and weighted network is a tree of that network consisting of all its nodes and the sum of weights of all its edges is minimum among all such possible spanning trees of the same network. In this study, we have developed a new GIS tool using most commonly known rudimentary algorithm called Prim's algorithm to construct the minimum spanning tree of a connected, undirected and weighted road network. This algorithm is based on the weight (adjacency) matrix of a weighted network and helps to solve complex network MST problem easily, efficiently and effectively. The selection of the appropriate algorithm is very essential otherwise it will be very hard to get an optimal result. In case of Road Transportation Network, it is very essential to find the optimal results by considering all the necessary points based on cost factor (time or distance). This paper is based on solving the Minimum Spanning Tree (MST) problem of a road network by finding it's minimum span by considering all the important network junction point. GIS technology is usually used to solve the network related problems like the optimal path problem, travelling salesman problem, vehicle routing problems, location-allocation problems etc. Therefore, in this study we have developed a customized GIS tool using Python script in ArcGIS software for the solution of MST problem for a Road Transportation Network of Dehradun city by considering distance and time as the impedance (cost) factors. It has a number of advantages like the users do not need a greater knowledge of the subject as the tool is user-friendly and that allows to access information varied and adapted the needs of the users. This GIS tool for MST can be applied for a nationwide plan called Prime Minister Gram Sadak Yojana in India to provide optimal all weather road connectivity to unconnected villages (points). This tool is also useful for constructing highways or railways spanning several
A direct algorithm for solution of incompressible three-dimensional unsteady Navier-Stokes equations
NASA Technical Reports Server (NTRS)
Osswald, G. A.; Ghia, K. N.; Ghia, U.
1987-01-01
A direct, implicit, numerical solution algorithm, with second-order accuracy in space and time, is constructed for the three-dimensional unsteady incompressible Navier-Stokes equations formulated in terms of velocity and vorticity, using generalized orthogonal coordinates to achieve the accurate solution of complex viscous flow configurations. A numerically stable, efficient, direct inversion procedure is developed for the computationally intensive divergence-curl elliptic velocity problem. This overdetermined partial differential operator is first formulated as a uniquely determined, nonsingular matrix-vector problem; this aspect of the procedure is a unique feature of the present analysis. The three-dimensional vorticity-transport equation is solved by a modified factorization technique which completely eliminates the need for any block-matrix inversions and only scalar tridiagonal matrices need to be inverted. The method is applied to the test problem of the three-dimensional flow within a shear-driven cubical box. Coherent streamwise vortex structures are observed within the steady-state flow at Re = 100.
NASA Technical Reports Server (NTRS)
Rogers, S. E.; Kwak, D.; Chang, J. L. C.
1986-01-01
The method of pseudocompressibility has been shown to be an efficient method for obtaining a steady-state solution to the incompressible Navier-Stokes equations. Recent improvements to this method include the use of a diagonal scheme for the inversion of the equations at each iteration. The necessary transformations have been derived for the pseudocompressibility equations in generalized coordinates. The diagonal algorithm reduces the computing time necessary to obtain a steady-state solution by a factor of nearly three. Implicit viscous terms are maintained in the equations, and it has become possible to use fourth-order implicit dissipation. The steady-state solution is unchanged by the approximations resulting from the diagonalization of the equations. Computed results for flow over a two-dimensional backward-facing step and a three-dimensional cylinder mounted normal to a flat plate are presented for both the old and new algorithms. The accuracy and computing efficiency of these algorithms are compared.
NASA Technical Reports Server (NTRS)
Boltz, F. W.
1984-01-01
An algorithm is presented for efficient p-iterative solution of the Lambert/Gauss orbit-determination problem using second-order Newton iteration. The algorithm is based on a universal transformation of Kepler's time-of-flight equation and approximate inverse solutions of this equation for short-way and long-way flight paths. The approximate solutions provide both good starting values for iteration and simplified computation of the second-order term in the iteration formula. Numerical results are presented which indicate that in many cases of practical significance (except those having collinear position vectors) the algorithm produces at least eight significant digits of accuracy with just two or three steps of iteration.
NASA Technical Reports Server (NTRS)
Leutenegger, Scott T.; Horton, Graham
1994-01-01
Recently the Multi-Level algorithm was introduced as a general purpose solver for the solution of steady state Markov chains. In this paper, we consider the performance of the Multi-Level algorithm for solving Nearly Completely Decomposable (NCD) Markov chains, for which special-purpose iteractive aggregation/disaggregation algorithms such as the Koury-McAllister-Stewart (KMS) method have been developed that can exploit the decomposability of the the Markov chain. We present experimental results indicating that the general-purpose Multi-Level algorithm is competitive, and can be significantly faster than the special-purpose KMS algorithm when Gauss-Seidel and Gaussian Elimination are used for solving the individual blocks.
Aerodynamic design optimization by using a continuous adjoint method
NASA Astrophysics Data System (ADS)
Luo, JiaQi; Xiong, JunTao; Liu, Feng
2014-07-01
This paper presents the fundamentals of a continuous adjoint method and the applications of this method to the aerodynamic design optimization of both external and internal flows. General formulation of the continuous adjoint equations and the corresponding boundary conditions are derived. With the adjoint method, the complete gradient information needed in the design optimization can be obtained by solving the governing flow equations and the corresponding adjoint equations only once for each cost function, regardless of the number of design parameters. An inverse design of airfoil is firstly performed to study the accuracy of the adjoint gradient and the effectiveness of the adjoint method as an inverse design method. Then the method is used to perform a series of single and multiple point design optimization problems involving the drag reduction of airfoil, wing, and wing-body configuration, and the aerodynamic performance improvement of turbine and compressor blade rows. The results demonstrate that the continuous adjoint method can efficiently and significantly improve the aerodynamic performance of the design in a shape optimization problem.
Ando, Tadashi; Chow, Edmond; Skolnick, Jeffrey
2013-01-01
Hydrodynamic interactions exert a critical effect on the dynamics of macromolecules. As the concentration of macromolecules increases, by analogy to the behavior of semidilute polymer solutions or the flow in porous media, one might expect hydrodynamic screening to occur. Hydrodynamic screening would have implications both for the understanding of macromolecular dynamics as well as practical implications for the simulation of concentrated macromolecular solutions, e.g., in cells. Stokesian dynamics (SD) is one of the most accurate methods for simulating the motions of N particles suspended in a viscous fluid at low Reynolds number, in that it considers both far-field and near-field hydrodynamic interactions. This algorithm traditionally involves an O(N3) operation to compute Brownian forces at each time step, although asymptotically faster but more complex SD methods are now available. Motivated by the idea of hydrodynamic screening, the far-field part of the hydrodynamic matrix in SD may be approximated by a diagonal matrix, which is equivalent to assuming that long range hydrodynamic interactions are completely screened. This approximation allows sparse matrix methods to be used, which can reduce the apparent computational scaling to O(N). Previously there were several simulation studies using this approximation for monodisperse suspensions. Here, we employ newly designed preconditioned iterative methods for both the computation of Brownian forces and the solution of linear systems, and consider the validity of this approximation in polydisperse suspensions. We evaluate the accuracy of the diagonal approximation method using an intracellular-like suspension. The diffusivities of particles obtained with this approximation are close to those with the original method. However, this approximation underestimates intermolecular correlated motions, which is a trade-off between accuracy and computing efficiency. The new method makes it possible to perform large-scale and
NASA Astrophysics Data System (ADS)
Ando, Tadashi; Chow, Edmond; Skolnick, Jeffrey
2013-09-01
Hydrodynamic interactions exert a critical effect on the dynamics of macromolecules. As the concentration of macromolecules increases, by analogy to the behavior of semidilute polymer solutions or the flow in porous media, one might expect hydrodynamic screening to occur. Hydrodynamic screening would have implications both for the understanding of macromolecular dynamics as well as practical implications for the simulation of concentrated macromolecular solutions, e.g., in cells. Stokesian dynamics (SD) is one of the most accurate methods for simulating the motions of N particles suspended in a viscous fluid at low Reynolds number, in that it considers both far-field and near-field hydrodynamic interactions. This algorithm traditionally involves an O(N3) operation to compute Brownian forces at each time step, although asymptotically faster but more complex SD methods are now available. Motivated by the idea of hydrodynamic screening, the far-field part of the hydrodynamic matrix in SD may be approximated by a diagonal matrix, which is equivalent to assuming that long range hydrodynamic interactions are completely screened. This approximation allows sparse matrix methods to be used, which can reduce the apparent computational scaling to O(N). Previously there were several simulation studies using this approximation for monodisperse suspensions. Here, we employ newly designed preconditioned iterative methods for both the computation of Brownian forces and the solution of linear systems, and consider the validity of this approximation in polydisperse suspensions. We evaluate the accuracy of the diagonal approximation method using an intracellular-like suspension. The diffusivities of particles obtained with this approximation are close to those with the original method. However, this approximation underestimates intermolecular correlated motions, which is a trade-off between accuracy and computing efficiency. The new method makes it possible to perform large-scale and
Adjoint active surfaces for localization and imaging.
Cook, Daniel A; Mueller, Martin Fritz; Fedele, Francesco; Yezzi, Anthony J
2015-01-01
This paper addresses the problem of localizing and segmenting regions embedded within a surrounding medium by characterizing their boundaries, as opposed to imaging the entirety of the volume. Active surfaces are used to directly reconstruct the shape of the region of interest. We describe the procedure for finding the optimal surface, which is computed iteratively via gradient descent that exploits the sensitivity of an error minimization functional to changes of the active surface. In doing so, we introduce the adjoint model to compute the sensitivity, and in this respect, the method shares common ground with several other disciplines, such as optimal control. Finally, we illustrate the proposed active surface technique in the framework of wave propagation governed by the scalar Helmholtz equation. Potential applications include electromagnetics, acoustics, geophysics, nondestructive testing, and medical imaging. PMID:25438311
Adjoint tomography of the southern California crust.
Tape, Carl; Liu, Qinya; Maggi, Alessia; Tromp, Jeroen
2009-08-21
Using an inversion strategy based on adjoint methods, we developed a three-dimensional seismological model of the southern California crust. The resulting model involved 16 tomographic iterations, which required 6800 wavefield simulations and a total of 0.8 million central processing unit hours. The new crustal model reveals strong heterogeneity, including local changes of +/-30% with respect to the initial three-dimensional model provided by the Southern California Earthquake Center. The model illuminates shallow features such as sedimentary basins and compositional contrasts across faults. It also reveals crustal features at depth that aid in the tectonic reconstruction of southern California, such as subduction-captured oceanic crustal fragments. The new model enables more realistic and accurate assessments of seismic hazard. PMID:19696349
Adjoints and Low-rank Covariance Representation
NASA Technical Reports Server (NTRS)
Tippett, Michael K.; Cohn, Stephen E.
2000-01-01
Quantitative measures of the uncertainty of Earth System estimates can be as important as the estimates themselves. Second moments of estimation errors are described by the covariance matrix, whose direct calculation is impractical when the number of degrees of freedom of the system state is large. Ensemble and reduced-state approaches to prediction and data assimilation replace full estimation error covariance matrices by low-rank approximations. The appropriateness of such approximations depends on the spectrum of the full error covariance matrix, whose calculation is also often impractical. Here we examine the situation where the error covariance is a linear transformation of a forcing error covariance. We use operator norms and adjoints to relate the appropriateness of low-rank representations to the conditioning of this transformation. The analysis is used to investigate low-rank representations of the steady-state response to random forcing of an idealized discrete-time dynamical system.
Self-adjointness and conservation laws of difference equations
NASA Astrophysics Data System (ADS)
Peng, Linyu
2015-06-01
A general theorem on conservation laws for arbitrary difference equations is proved. The theorem is based on an introduction of an adjoint system related with a given difference system, and it does not require the existence of a difference Lagrangian. It is proved that the system, combined by the original system and its adjoint system, is governed by a variational principle, which inherits all symmetries of the original system. Noether's theorem can then be applied. With some special techniques, e.g. self-adjointness properties, this allows us to obtain conservation laws for difference equations, which are not necessary governed by Lagrangian formalisms.
A forward operator and its adjoint for GPS slant total delays
NASA Astrophysics Data System (ADS)
Zus, Florian; Dick, Galina; Heise, Stefan; Wickert, Jens
2015-05-01
In a recent study we developed a fast and accurate algorithm to compute Global Positioning System (GPS) Slant Total Delay (STDs) utilizing numerical weather model data. Having developed a forward operator we construct in this study the tangent linear (adjoint) operator by application of the chain rule of differential calculus in forward (reverse) mode. Armed with these operators we show in a simulation study the potential benefit of GPS STDs in inverse modeling. We conclude that the developed operators are tailored for three (four)-dimensional variational data assimilation and/or travel time tomography.
NASA Astrophysics Data System (ADS)
Shi, Lei; Wang, Z. J.
2015-08-01
Adjoint-based mesh adaptive methods are capable of distributing computational resources to areas which are important for predicting an engineering output. In this paper, we develop an adjoint-based h-adaptation approach based on the high-order correction procedure via reconstruction formulation (CPR) to minimize the output or functional error. A dual-consistent CPR formulation of hyperbolic conservation laws is developed and its dual consistency is analyzed. Super-convergent functional and error estimate for the output with the CPR method are obtained. Factors affecting the dual consistency, such as the solution point distribution, correction functions, boundary conditions and the discretization approach for the non-linear flux divergence term, are studied. The presented method is then used to perform simulations for the 2D Euler and Navier-Stokes equations with mesh adaptation driven by the adjoint-based error estimate. Several numerical examples demonstrate the ability of the presented method to dramatically reduce the computational cost comparing with uniform grid refinement.
NASA Technical Reports Server (NTRS)
Palmer, Grant; Venkatapathy, Ethiraj
1993-01-01
Three solution algorithms, explicit underrelaxation, point implicit, and lower upper symmetric Gauss-Seidel (LUSGS), are used to compute nonequilibrium flow around the Apollo 4 return capsule at 62 km altitude. By varying the Mach number, the efficiency and robustness of the solution algorithms were tested for different levels of chemical stiffness. The performance of the solution algorithms degraded as the Mach number and stiffness of the flow increased. At Mach 15, 23, and 30, the LUSGS method produces an eight order of magnitude drop in the L2 norm of the energy residual in 1/3 to 1/2 the Cray C-90 computer time as compared to the point implicit and explicit under-relaxation methods. The explicit under-relaxation algorithm experienced convergence difficulties at Mach 23 and above. At Mach 40 the performance of the LUSGS algorithm deteriorates to the point it is out-performed by the point implicit method. The effects of the viscous terms are investigated. Grid dependency questions are explored.
Sensitivity of Lumped Constraints Using the Adjoint Method
NASA Technical Reports Server (NTRS)
Akgun, Mehmet A.; Haftka, Raphael T.; Wu, K. Chauncey; Walsh, Joanne L.
1999-01-01
Adjoint sensitivity calculation of stress, buckling and displacement constraints may be much less expensive than direct sensitivity calculation when the number of load cases is large. Adjoint stress and displacement sensitivities are available in the literature. Expressions for local buckling sensitivity of isotropic plate elements are derived in this study. Computational efficiency of the adjoint method is sensitive to the number of constraints and, therefore, the method benefits from constraint lumping. A continuum version of the Kreisselmeier-Steinhauser (KS) function is chosen to lump constraints. The adjoint and direct methods are compared for three examples: a truss structure, a simple HSCT wing model, and a large HSCT model. These sensitivity derivatives are then used in optimization.
NASA Technical Reports Server (NTRS)
Pflaum, Christoph
1996-01-01
A multilevel algorithm is presented that solves general second order elliptic partial differential equations on adaptive sparse grids. The multilevel algorithm consists of several V-cycles. Suitable discretizations provide that the discrete equation system can be solved in an efficient way. Numerical experiments show a convergence rate of order Omicron(1) for the multilevel algorithm.
NASA Technical Reports Server (NTRS)
Chima, R. V.; Johnson, G. M.
1983-01-01
A multiple-grid algorithm for use in efficiently obtaining steady solutions to the Euler and Navier-Stokes equations is presented. The convergence of the explicit MacCormack algorithm on a fine grid is accelerated by propagating transients from the domain using a sequence of successively coarser grids. Both the fine and coarse grid schemes are readily vectorizable. The combination of multiple-gridding and vectorization results in substantially reduced computational times for the numerical solution of a wide range of flow problems. Results are presented for subsonic, transonic, and supersonic inviscid flows and for subsonic attached and separated laminar viscous flows. Work reduction factors over a scalar, single-grid algorithm range as high as 76.8. Previously announced in STAR as N83-24467
Jenny, Patrick Torrilhon, Manuel; Heinz, Stefan
2010-02-20
In this paper, a stochastic model is presented to simulate the flow of gases, which are not in thermodynamic equilibrium, like in rarefied or micro situations. For the interaction of a particle with others, statistical moments of the local ensemble have to be evaluated, but unlike in molecular dynamics simulations or DSMC, no collisions between computational particles are considered. In addition, a novel integration technique allows for time steps independent of the stochastic time scale. The stochastic model represents a Fokker-Planck equation in the kinetic description, which can be viewed as an approximation to the Boltzmann equation. This allows for a rigorous investigation of the relation between the new model and classical fluid and kinetic equations. The fluid dynamic equations of Navier-Stokes and Fourier are fully recovered for small relaxation times, while for larger values the new model extents into the kinetic regime. Numerical studies demonstrate that the stochastic model is consistent with Navier-Stokes in that limit, but also that the results become significantly different, if the conditions for equilibrium are invalid. The application to the Knudsen paradox demonstrates the correctness and relevance of this development, and comparisons with existing kinetic equations and standard solution algorithms reveal its advantages. Moreover, results of a test case with geometrically complex boundaries are presented.
Surface wave sensitivity: mode summation versus adjoint SEM
NASA Astrophysics Data System (ADS)
Zhou, Ying; Liu, Qinya; Tromp, Jeroen
2011-12-01
We compare finite-frequency phase and amplitude sensitivity kernels calculated based on frequency-domain surface wave mode summation and a time-domain adjoint method. The adjoint calculations involve a forward wavefield generated by an earthquake and an adjoint wavefield generated at a seismic receiver. We determine adjoint sources corresponding to frequency-dependent phase and amplitude measurements made using a multitaper technique, which may be applied to any single-taper measurement, including box car windowing. We calculate phase and amplitude sensitivity kernels using an adjoint method based on wave propagation simulations using a spectral element method (SEM). Sensitivity kernels calculated using the adjoint SEM are in good agreement with kernels calculated based on mode summation. In general, the adjoint SEM is more computationally expensive than mode summation in global studies. The advantage of the adjoint SEM lies in the calculation of sensitivity kernels in 3-D earth models. We compare surface wave sensitivity kernels computed in 1-D and 3-D reference earth models and show that (1) lateral wave speed heterogeneities may affect the geometry and amplitude of surface wave sensitivity; (2) sensitivity kernels of long-period surface waves calculated in 1-D model PREM and 3-D models S20RTS+CRUST2.0 and FFSW1+CRUST2.0 do not show significant differences, indicating that the use of a 1-D reference model is adequate in global inversions of long-period surface waves (periods of 50 s and longer); and (3) the differences become significant for short-period Love waves when mode coupling is sensitive to large differences in reference crustal structure. Finally, we show that sensitivity kernels in anelastic earth models may be calculated in purely elastic earth models provided physical dispersion is properly accounted for.
Adjoint Function: Physical Basis of Variational & Perturbation Theory in Transport
Energy Science and Technology Software Center (ESTSC)
2009-07-27
Version 00 Dr. J.D. Lewins has now released the following legacy book for free distribution: Importance: The Adjoint Function: The Physical Basis of Variational and Perturbation Theory in Transport and Diffusion Problems, North-Holland Publishing Company - Amsterdam, 582 pages, 1966 Introduction: Continuous Systems and the Variational Principle 1. The Fundamental Variational Principle 2. The Importance Function 3. Adjoint Equations 4. Variational Methods 5. Perturbation and Iterative Methods 6. Non-Linear Theory
A family of solution algorithms for nonlinear structural analysis based on relaxation equations
NASA Technical Reports Server (NTRS)
Park, K. C.
1981-01-01
A family of hierarchical algorithms for nonlinear structural equations are presented. The algorithms are based on the Davidenko-Branin type homotopy and shown to yield consistent hierarchical perturbation equations. The algorithms appear to be particularly suitable to problems involving bifurcation and limit point calculations. An important by-product of the algorithms is that it provides a systematic and economical means for computing the stepsize at each iteration stage when a Newton-like method is employed to solve the systems of equations. Some sample problems are provided to illustrate the characteristics of the algorithms.
Universal Racah matrices and adjoint knot polynomials: Arborescent knots
NASA Astrophysics Data System (ADS)
Mironov, A.; Morozov, A.
2016-04-01
By now it is well established that the quantum dimensions of descendants of the adjoint representation can be described in a universal form, independent of a particular family of simple Lie algebras. The Rosso-Jones formula then implies a universal description of the adjoint knot polynomials for torus knots, which in particular unifies the HOMFLY (SUN) and Kauffman (SON) polynomials. For E8 the adjoint representation is also fundamental. We suggest to extend the universality from the dimensions to the Racah matrices and this immediately produces a unified description of the adjoint knot polynomials for all arborescent (double-fat) knots, including twist, 2-bridge and pretzel. Technically we develop together the universality and the "eigenvalue conjecture", which expresses the Racah and mixing matrices through the eigenvalues of the quantum R-matrix, and for dealing with the adjoint polynomials one has to extend it to the previously unknown 6 × 6 case. The adjoint polynomials do not distinguish between mutants and therefore are not very efficient in knot theory, however, universal polynomials in higher representations can probably be better in this respect.
Comparison of adjoint and nudging methods to initialise ice sheet model basal conditions
NASA Astrophysics Data System (ADS)
Mosbeux, Cyrille; Gillet-Chaulet, Fabien; Gagliardini, Olivier
2016-07-01
Ice flow models are now routinely used to forecast the ice sheets' contribution to 21st century sea-level rise. For such short term simulations, the model response is greatly affected by the initial conditions. Data assimilation algorithms have been developed to invert for the friction of the ice on its bedrock using observed surface velocities. A drawback of these methods is that remaining uncertainties, especially in the bedrock elevation, lead to non-physical ice flux divergence anomalies resulting in undesirable transient effects. In this study, we compare two different assimilation algorithms based on adjoints and nudging to constrain both bedrock friction and elevation. Using synthetic twin experiments with realistic observation errors, we show that the two algorithms lead to similar performances in reconstructing both variables and allow the flux divergence anomalies to be significantly reduced.
Periodic differential equations with self-adjoint monodromy operator
NASA Astrophysics Data System (ADS)
Yudovich, V. I.
2001-04-01
A linear differential equation \\dot u=A(t)u with p-periodic (generally speaking, unbounded) operator coefficient in a Euclidean or a Hilbert space \\mathbb H is considered. It is proved under natural constraints that the monodromy operator U_p is self-adjoint and strictly positive if A^*(-t)=A(t) for all t\\in\\mathbb R.It is shown that Hamiltonian systems in the class under consideration are usually unstable and, if they are stable, then the operator U_p reduces to the identity and all solutions are p-periodic.For higher frequencies averaged equations are derived. Remarkably, high-frequency modulation may double the number of critical values.General results are applied to rotational flows with cylindrical components of the velocity a_r=a_z=0, a_\\theta=\\lambda c(t)r^\\beta, \\beta<-1, c(t) is an even p-periodic function, and also to several problems of free gravitational convection of fluids in periodic fields.
Parallelized Three-Dimensional Resistivity Inversion Using Finite Elements And Adjoint State Methods
NASA Astrophysics Data System (ADS)
Schaa, Ralf; Gross, Lutz; Du Plessis, Jaco
2015-04-01
The resistivity method is one of the oldest geophysical exploration methods, which employs one pair of electrodes to inject current into the ground and one or more pairs of electrodes to measure the electrical potential difference. The potential difference is a non-linear function of the subsurface resistivity distribution described by an elliptic partial differential equation (PDE) of the Poisson type. Inversion of measured potentials solves for the subsurface resistivity represented by PDE coefficients. With increasing advances in multichannel resistivity acquisition systems (systems with more than 60 channels and full waveform recording are now emerging), inversion software require efficient storage and solver algorithms. We developed the finite element solver Escript, which provides a user-friendly programming environment in Python to solve large-scale PDE-based problems (see https://launchpad.net/escript-finley). Using finite elements, highly irregular shaped geology and topography can readily be taken into account. For the 3D resistivity problem, we have implemented the secondary potential approach, where the PDE is decomposed into a primary potential caused by the source current and the secondary potential caused by changes in subsurface resistivity. The primary potential is calculated analytically, and the boundary value problem for the secondary potential is solved using nodal finite elements. This approach removes the singularity caused by the source currents and provides more accurate 3D resistivity models. To solve the inversion problem we apply a 'first optimize then discretize' approach using the quasi-Newton scheme in form of the limited-memory Broyden-Fletcher-Goldfarb-Shanno (L-BFGS) method (see Gross & Kemp 2013). The evaluation of the cost function requires the solution of the secondary potential PDE for each source current and the solution of the corresponding adjoint-state PDE for the cost function gradients with respect to the subsurface
NASA Astrophysics Data System (ADS)
Kocaogul, Ibrahim; Hu, Fang; Li, Xiaodong
2014-03-01
Radiation of acoustic waves at all frequencies can be obtained by Time Domain Wave Packet (TDWP) method in a single time domain computation. Other benefit of the TDWP method is that it makes possible the separation of acoustic and instability wave in the shear flow. The TDWP method is also particularly useful for computations in the ducted or waveguide environments where incident wave modes can be imposed cleanly without a potentially long transient period. The adjoint equations for the linearized Euler equations are formulated for the Cartesian coordinates. Analytical solution for adjoint equations is derived by using Green's function in 2D and 3D. The derivation of reciprocal relations is presented for closed and open ducts. The adjoint equations are then solved numerically in reversed time by the TDWP method. Reciprocal relation between the duct mode amplitudes and far field point sources in the presence of the exhaust shear flow is computed and confirmed numerically. Applications of the adjoint problem to closed and open ducts are also presented.
Baryogenesis via leptogenesis in adjoint SU(5)
Blanchet, Steve; Fileviez Perez, Pavel E-mail: fileviez@physics.wisc.edu
2008-08-15
The possibility of explaining the baryon asymmetry in the Universe through the leptogenesis mechanism in the context of adjoint SU(5) is investigated. In this model neutrino masses are generated through the type I and type III seesaw mechanisms, and the field responsible for the type III seesaw, called {rho}{sub 3}, generates the B-L asymmetry needed to satisfy the observed value of the baryon asymmetry in the Universe. We find that the CP asymmetry originates only from the vertex correction, since the self-energy contribution is not present. When neutrino masses have a normal hierarchy, successful leptogenesis is possible for 10{sup 11} GeV{approx}
Hybrid algorithm: A cost efficient solution for ONU placement in Fiber-Wireless (FiWi) network
NASA Astrophysics Data System (ADS)
Bhatt, Uma Rathore; Chouhan, Nitin; Upadhyay, Raksha
2015-03-01
Fiber-Wireless (FiWi) network is a promising access technology as it integrates the technical merits of optical and wireless access networks. FiWi provides large bandwidth and high stability of optical network and lower cost of wireless network respectively. Therefore, FiWi gives users to access broadband services in an "anywhere-anytime" way. One of the key issues in FiWi network is its deployment cost, which depends on the number of ONUs in the network. Therefore optimal placement of ONUs is desirable to design a cost effective network. In this paper, we propose an algorithm for optimal placement of ONUs. First we place an ONU in the center of each grid then we form a set of wireless routers associated with each ONU according to wireless hop number. The number of ONUs are minimized in such a way, that all the wireless routers can communicate to at least one of the ONUs. The number of ONUs in the network further reduced by using genetic algorithm. The effectiveness of the proposed algorithm is tested by considering Internet traffic as well as peer-to-peer (p2p) traffic in the network, which is a current need. Simulation results show that the proposed algorithm is better than existing algorithms in minimizing number of ONUs in the network for both types of traffics. Hence proposed algorithm offers cost effective solution to design the FiWi network.
Bouallègue, Fayçal Ben; Crouzet, Jean-François; Comtat, Claude; Fourcade, Marjolaine; Mohammadi, Bijan; Mariano-Goulart, Denis
2007-07-01
This paper presents an extended 3-D exact rebinning formula in the Fourier space that leads to an iterative reprojection algorithm (iterative FOREPROJ), which enables the estimation of unmeasured oblique projection data on the basis of the whole set of measured data. In first approximation, this analytical formula also leads to an extended Fourier rebinning equation that is the basis for an approximate reprojection algorithm (extended FORE). These algorithms were evaluated on numerically simulated 3-D positron emission tomography (PET) data for the solution of the truncation problem, i.e., the estimation of the missing portions in the oblique projection data, before the application of algorithms that require complete projection data such as some rebinning methods (FOREX) or 3-D reconstruction algorithms (3DRP or direct Fourier methods). By taking advantage of all the 3-D data statistics, the iterative FOREPROJ reprojection provides a reliable alternative to the classical FOREPROJ method, which only exploits the low-statistics nonoblique data. It significantly improves the quality of the external reconstructed slices without loss of spatial resolution. As for the approximate extended FORE algorithm, it clearly exhibits limitations due to axial interpolations, but will require clinical studies with more realistic measured data in order to decide on its pertinence. PMID:17649913
NASA Astrophysics Data System (ADS)
Hunziker, J.; Thorbecke, J.; Slob, E. C.
2014-12-01
Commonly, electromagnetic measurements for exploring and monitoring hydrocarbon reservoirs are inverted for the subsurface conductivity distribution by minimizing the difference between the actual data and a forward modeled dataset. The convergence of the inversion process to the correct solution strongly depends on the shape of the solution space. Since this is a non-linear problem, there exist a multitude of minima of which only the global one provides the correct conductivity values. To easily find the global minimum we desire it to have a broad cone of attraction, while it should also feature a very narrow bottom in order to obtain the subsurface conductivity with high resolution. In this study, we aim to determine which combination of input data corresponds to a favorable shape of the solution space. Since the solution space is N-dimensional, with N being the number of unknown subsurface parameters, plotting it is out of the question. In our approach, we use a genetic algorithm (Goldberg, 1989) to probe the solution space. Such algorithms have the advantage that every run of the same problem will end up at a different solution. Most of these solutions are expected to lie close to the global minimum. A situation where only few runs end up in the global minimum indicates that the solution space consists of a lot of local minima or that the cone of attraction of the global minimum is small. If a lot of runs end up with a similar data-misfit but with a large spread of the subsurface medium parameters in one or more direction, it can be concluded that the chosen data-input is not sensitive with respect to that direction. Compared to the study of Hunziker et al. 2014, we allow also to invert for subsurface boundaries and include more combinations of input datasets. The results so far suggest that it is essential to include the magnetic field in the inversion process in order to find the anisotropic conductivity values. ReferencesGoldberg, D. E., 1989. Genetic
Diffusion Acceleration Schemes for Self-Adjoint Angular Flux Formulation with a Void Treatment
Yaqi Wang; Hongbin Zhang; Richard C. Martineau
2014-02-01
A Galerkin weak form for the monoenergetic neutron transport equation with a continuous finite element method and discrete ordinate method is developed based on self-adjoint angular flux formulation. This weak form is modified for treating void regions. A consistent diffusion scheme is developed with projection. Correction terms of the diffusion scheme are derived to reproduce the transport scalar flux. A source iteration that decouples the solution of all directions with both linear and nonlinear diffusion accelerations is developed and demonstrated. One-dimensional Fourier analysis is conducted to demonstrate the stability of the linear and nonlinear diffusion accelerations. Numerical results of these schemes are presented.
Receptivity in parallel flows: An adjoint approach
NASA Technical Reports Server (NTRS)
Hill, D. Christopher
1993-01-01
Linear receptivity studies in parallel flows are aimed at understanding how external forcing couples to the natural unstable motions which a flow can support. The vibrating ribbon problem models the original Schubauer and Skramstad boundary layer experiment and represents the classic boundary layer receptivity problem. The process by which disturbances are initiated in convectively-unstable jets and shear layers has also received attention. Gaster was the first to handle the boundary layer analysis with the recognition that spatial modes, rather than temporal modes, were relevant when studying convectively-unstable flows that are driven by a time-harmonic source. The amplitude of the least stable spatial mode, far downstream of the source, is related to the source strength by a coupling coefficient. The determination of this coefficient is at the heart of this type of linear receptivity study. The first objective of the present study was to determine whether the various wave number derivative factors, appearing in the coupling coefficients for linear receptivity problems, could be reexpressed in a simpler form involving adjoint eigensolutions. Secondly, it was hoped that the general nature of this simplification could be shown; indeed, a rather elegant characterization of the receptivity properties of spatial instabilities does emerge. The analysis is quite distinct from the usual Fourier-inversion procedures, although a detailed knowledge of the spectrum of the Orr-Sommerfeld equation is still required. Since the cylinder wake analysis proved very useful in addressing control considerations, the final objective was to provide a foundation upon which boundary layer control theory may be developed.
Adjoint simulation of stream depletion due to aquifer pumping.
Neupauer, Roseanna M; Griebling, Scott A
2012-01-01
If an aquifer is hydraulically connected to an adjacent stream, a pumping well operating in the aquifer will draw some water from aquifer storage and some water from the stream, causing stream depletion. Several analytical, semi-analytical, and numerical approaches have been developed to estimate stream depletion due to pumping. These approaches are effective if the well location is known. If a new well is to be installed, it may be desirable to install the well at a location where stream depletion is minimal. If several possible locations are considered for the location of a new well, stream depletion would have to be estimated for all possible well locations, which can be computationally inefficient. The adjoint approach for estimating stream depletion is a more efficient alternative because with one simulation of the adjoint model, stream depletion can be estimated for pumping at a well at any location. We derive the adjoint equations for a coupled system with a confined aquifer, an overlying unconfined aquifer, and a river that is hydraulically connected to the unconfined aquifer. We assume that the stage in the river is known, and is independent of the stream depletion, consistent with the assumptions of the MODFLOW river package. We describe how the adjoint equations can be solved using MODFLOW. In an illustrative example, we show that for this scenario, the adjoint approach is as accurate as standard forward numerical simulation methods, and requires substantially less computational effort. PMID:22182421
Adjoint-Based Methodology for Time-Dependent Optimization
NASA Technical Reports Server (NTRS)
Yamaleev, N. K.; Diskin, B.; Nielsen, E. J.
2008-01-01
This paper presents a discrete adjoint method for a broad class of time-dependent optimization problems. The time-dependent adjoint equations are derived in terms of the discrete residual of an arbitrary finite volume scheme which approximates unsteady conservation law equations. Although only the 2-D unsteady Euler equations are considered in the present analysis, this time-dependent adjoint method is applicable to the 3-D unsteady Reynolds-averaged Navier-Stokes equations with minor modifications. The discrete adjoint operators involving the derivatives of the discrete residual and the cost functional with respect to the flow variables are computed using a complex-variable approach, which provides discrete consistency and drastically reduces the implementation and debugging cycle. The implementation of the time-dependent adjoint method is validated by comparing the sensitivity derivative with that obtained by forward mode differentiation. Our numerical results show that O(10) optimization iterations of the steepest descent method are needed to reduce the objective functional by 3-6 orders of magnitude for test problems considered.
Frolov, Vladimir; Backhaus, Scott N.; Chertkov, Michael
2014-01-14
In a companion manuscript, we developed a novel optimization method for placement, sizing, and operation of Flexible Alternating Current Transmission System (FACTS) devices to relieve transmission network congestion. Specifically, we addressed FACTS that provide Series Compensation (SC) via modification of line inductance. In this manuscript, this heuristic algorithm and its solutions are explored on a number of test cases: a 30-bus test network and a realistically-sized model of the Polish grid (~2700 nodes and ~3300 lines). The results on the 30-bus network are used to study the general properties of the solutions including non-locality and sparsity. The Polish grid is used as a demonstration of the computational efficiency of the heuristics that leverages sequential linearization of power flow constraints and cutting plane methods that take advantage of the sparse nature of the SC placement solutions. Using these approaches, the algorithm is able to solve an instance of Polish grid in tens of seconds. We explore the utility of the algorithm by analyzing transmission networks congested by (a) uniform load growth, (b) multiple overloaded configurations, and (c) sequential generator retirements
Frolov, Vladimir; Backhaus, Scott; Chertkov, Misha
2014-10-01
In a companion manuscript, we developed a novel optimization method for placement, sizing, and operation of Flexible Alternating Current Transmission System (FACTS) devices to relieve transmission network congestion. Specifically, we addressed FACTS that provide Series Compensation (SC) via modification of line inductance. In this manuscript, this heuristic algorithm and its solutions are explored on a number of test cases: a 30-bus test network and a realistically-sized model of the Polish grid (~ 2700 nodes and ~ 3300 lines). The results on the 30-bus network are used to study the general properties of the solutions including non-locality and sparsity. The Polish grid is used as a demonstration of the computational efficiency of the heuristics that leverages sequential linearization of power flow constraints and cutting plane methods that take advantage of the sparse nature of the SC placement solutions. Using these approaches, the algorithm is able to solve an instance of Polish grid in tens of seconds. We explore the utility of the algorithm by analyzing transmission networks congested by (a) uniform load growth, (b) multiple overloaded configurations, and (c) sequential generator retirements.
NASA Astrophysics Data System (ADS)
Frolov, Vladimir; Backhaus, Scott; Chertkov, Misha
2014-10-01
In a companion manuscript (Frolov et al 2014 New J. Phys. 16 art. no.) , we developed a novel optimization method for the placement, sizing, and operation of flexible alternating current transmission system (FACTS) devices to relieve transmission network congestion. Specifically, we addressed FACTS that provide series compensation (SC) via modification of line inductance. In this sequel manuscript, this heuristic algorithm and its solutions are explored on a number of test cases: a 30-bus test network and a realistically-sized model of the Polish grid (˜2700 nodes and ˜3300 lines). The results from the 30-bus network are used to study the general properties of the solutions, including nonlocality and sparsity. The Polish grid is used to demonstrate the computational efficiency of the heuristics that leverage sequential linearization of power flow constraints, and cutting plane methods that take advantage of the sparse nature of the SC placement solutions. Using these approaches, we can use the algorithm to solve a Polish transmission grid in tens of seconds. We explore the utility of the algorithm by analyzing transmission networks congested by (i) uniform load growth, (ii) multiple overloaded configurations, and (iii) sequential generator retirements.
Frolov, Vladimir; Backhaus, Scott; Chertkov, Misha
2014-10-01
In a companion manuscript, we developed a novel optimization method for placement, sizing, and operation of Flexible Alternating Current Transmission System (FACTS) devices to relieve transmission network congestion. Specifically, we addressed FACTS that provide Series Compensation (SC) via modification of line inductance. In this manuscript, this heuristic algorithm and its solutions are explored on a number of test cases: a 30-bus test network and a realistically-sized model of the Polish grid (~ 2700 nodes and ~ 3300 lines). The results on the 30-bus network are used to study the general properties of the solutions including non-locality and sparsity. The Polishmore » grid is used as a demonstration of the computational efficiency of the heuristics that leverages sequential linearization of power flow constraints and cutting plane methods that take advantage of the sparse nature of the SC placement solutions. Using these approaches, the algorithm is able to solve an instance of Polish grid in tens of seconds. We explore the utility of the algorithm by analyzing transmission networks congested by (a) uniform load growth, (b) multiple overloaded configurations, and (c) sequential generator retirements.« less
Marble Algorithm: a solution to estimating ecological niches from presence-only records
Qiao, Huijie; Lin, Congtian; Jiang, Zhigang; Ji, Liqiang
2015-01-01
We describe an algorithm that helps to predict potential distributional areas for species using presence-only records. The Marble Algorithm is a density-based clustering program based on Hutchinson’s concept of ecological niches as multidimensional hypervolumes in environmental space. The algorithm characterizes this niche space using the density-based spatial clustering of applications with noise (DBSCAN) algorithm. When MA is provided with a set of occurrence points in environmental space, the algorithm determines two parameters that allow the points to be grouped into several clusters. These clusters are used as reference sets describing the ecological niche, which can then be mapped onto geographic space and used as the potential distribution of the species. We used both virtual species and ten empirical datasets to compare MA with other distribution-modeling tools, including Bioclimate Analysis and Prediction System, Environmental Niche Factor Analysis, the Genetic Algorithm for Rule-set Production, Maximum Entropy Modeling, Artificial Neural Networks, Climate Space Models, Classification Tree Analysis, Generalised Additive Models, Generalised Boosted Models, Generalised Linear Models, Multivariate Adaptive Regression Splines and Random Forests. Results indicate that MA predicts potential distributional areas with high accuracy, moderate robustness, and above-average transferability on all datasets, particularly when dealing with small numbers of occurrences. PMID:26387771
Adjoint sensitivity analysis of plasmonic structures using the FDTD method.
Zhang, Yu; Ahmed, Osman S; Bakr, Mohamed H
2014-05-15
We present an adjoint variable method for estimating the sensitivities of arbitrary responses with respect to the parameters of dispersive discontinuities in nanoplasmonic devices. Our theory is formulated in terms of the electric field components at the vicinity of perturbed discontinuities. The adjoint sensitivities are computed using at most one extra finite-difference time-domain (FDTD) simulation regardless of the number of parameters. Our approach is illustrated through the sensitivity analysis of an add-drop coupler consisting of a square ring resonator between two parallel waveguides. The computed adjoint sensitivities of the scattering parameters are compared with those obtained using the accurate but computationally expensive central finite difference approach. PMID:24978258
Sonic Boom Mitigation Through Aircraft Design and Adjoint Methodology
NASA Technical Reports Server (NTRS)
Rallabhandi, Siriam K.; Diskin, Boris; Nielsen, Eric J.
2012-01-01
This paper presents a novel approach to design of the supersonic aircraft outer mold line (OML) by optimizing the A-weighted loudness of sonic boom signature predicted on the ground. The optimization process uses the sensitivity information obtained by coupling the discrete adjoint formulations for the augmented Burgers Equation and Computational Fluid Dynamics (CFD) equations. This coupled formulation links the loudness of the ground boom signature to the aircraft geometry thus allowing efficient shape optimization for the purpose of minimizing the impact of loudness. The accuracy of the adjoint-based sensitivities is verified against sensitivities obtained using an independent complex-variable approach. The adjoint based optimization methodology is applied to a configuration previously optimized using alternative state of the art optimization methods and produces additional loudness reduction. The results of the optimizations are reported and discussed.
Design of an FPGA-Based Algorithm for Real-Time Solutions of Statistics-Based Positioning.
Dewitt, Don; Johnson-Williams, Nathan G; Miyaoka, Robert S; Li, Xiaoli; Lockhart, Cate; Lewellen, Tom K; Hauck, Scott
2010-02-01
We report on the implementation of an algorithm and hardware platform to allow real-time processing of the statistics-based positioning (SBP) method for continuous miniature crystal element (cMiCE) detectors. The SBP method allows an intrinsic spatial resolution of ~1.6 mm FWHM to be achieved using our cMiCE design. Previous SBP solutions have required a postprocessing procedure due to the computation and memory intensive nature of SBP. This new implementation takes advantage of a combination of algebraic simplifications, conversion to fixed-point math, and a hierarchal search technique to greatly accelerate the algorithm. For the presented seven stage, 127 × 127 bin LUT implementation, these algorithm improvements result in a reduction from >7 × 10(6) floating-point operations per event for an exhaustive search to < 5 × 10(3) integer operations per event. Simulations show nearly identical FWHM positioning resolution for this accelerated SBP solution, and positioning differences of <0.1 mm from the exhaustive search solution. A pipelined field programmable gate array (FPGA) implementation of this optimized algorithm is able to process events in excess of 250 K events per second, which is greater than the maximum expected coincidence rate for an individual detector. In contrast with all detectors being processed at a centralized host, as in the current system, a separate FPGA is available at each detector, thus dividing the computational load. These methods allow SBP results to be calculated in real-time and to be presented to the image generation components in real-time. A hardware implementation has been developed using a commercially available prototype board. PMID:21197135
Design of an FPGA-Based Algorithm for Real-Time Solutions of Statistics-Based Positioning
DeWitt, Don; Johnson-Williams, Nathan G.; Miyaoka, Robert S.; Li, Xiaoli; Lockhart, Cate; Lewellen, Tom K.; Hauck, Scott
2010-01-01
We report on the implementation of an algorithm and hardware platform to allow real-time processing of the statistics-based positioning (SBP) method for continuous miniature crystal element (cMiCE) detectors. The SBP method allows an intrinsic spatial resolution of ~1.6 mm FWHM to be achieved using our cMiCE design. Previous SBP solutions have required a postprocessing procedure due to the computation and memory intensive nature of SBP. This new implementation takes advantage of a combination of algebraic simplifications, conversion to fixed-point math, and a hierarchal search technique to greatly accelerate the algorithm. For the presented seven stage, 127 × 127 bin LUT implementation, these algorithm improvements result in a reduction from >7 × 106 floating-point operations per event for an exhaustive search to < 5 × 103 integer operations per event. Simulations show nearly identical FWHM positioning resolution for this accelerated SBP solution, and positioning differences of <0.1 mm from the exhaustive search solution. A pipelined field programmable gate array (FPGA) implementation of this optimized algorithm is able to process events in excess of 250 K events per second, which is greater than the maximum expected coincidence rate for an individual detector. In contrast with all detectors being processed at a centralized host, as in the current system, a separate FPGA is available at each detector, thus dividing the computational load. These methods allow SBP results to be calculated in real-time and to be presented to the image generation components in real-time. A hardware implementation has been developed using a commercially available prototype board. PMID:21197135
NASA Technical Reports Server (NTRS)
Fatemi, Emad; Jerome, Joseph; Osher, Stanley
1989-01-01
A micron n+ - n - n+ silicon diode is simulated via the hydrodynamic model for carrier transport. The numerical algorithms employed are for the non-steady case, and a limiting process is used to reach steady state. The simulation employs shock capturing algorithms, and indeed shocks, or very rapid transition regimes, are observed in the transient case for the coupled system, consisting of the potential equation and the conservation equations describing charge, momentum, and energy transfer for the electron carriers. These algorithms, termed essentially non-oscillatory, were successfully applied in other contexts to model the flow in gas dynamics, magnetohydrodynamics, and other physical situations involving the conservation laws in fluid mechanics. The method here is first order in time, but the use of small time steps allows for good accuracy. Runge-Kutta methods allow one to achieve higher accuracy in time if desired. The spatial accuracy is of high order in regions of smoothness.
Reconstruction of ocean circulation from sparse data using the adjoint method: LGM and the present
NASA Astrophysics Data System (ADS)
Kurahashi-Nakamura, T.; Losch, M. J.; Paul, A.; Mulitza, S.; Schulz, M.
2010-12-01
Understanding the behavior of the Earth's climate system under different conditions in the past is the basis for more robust projections of future climate. It is thought that the ocean circulation plays a very important role in the climate system, because it can greatly affect climate by dynamic-thermodynamic (as a medium of heat transport) and biogeochemical processes (by affecting the global carbon cycle). In this context, studying the period of the Last Glacial Maximum (LGM) is particularly promising, as it represents a climate state that is very different from today. Furthermore the LGM, compared to other paleoperiods, is characterized by a relatively good paleo-data coverage. Unfortunately, the ocean circulation during the LGM is still uncertain, with a range of climate models estimating both a stronger and a weaker formation rate of North Atlantic Deep Water (NADW) as compared to the present rate. Here, we present a project aiming at reducing this uncertainty by combining proxy data with a numerical ocean model using variational techniques. Our approach, the so-called adjoint method, employs a quadratic cost function of model-data differences weighted by their prior error estimates. We seek an optimal state estimate at the global minimum of the cost function by varying the independent control variables such as initial conditions (e.g. temperature), boundary conditions (e.g. surface winds, heat flux), or internal parameters (e.g. vertical diffusivity). The adjoint or dual model computes the gradient of the cost function with respect to these control variables and thus provides the information required by gradient descent algorithms. The gradients themselves provide valuable information about the sensitivity of the system to perturbations in the control variables. We use the Massachusetts Institute of Technology ocean general circulation model (MITgcm) with a cubed-sphere grid system that avoids converging grid lines and pole singularities. This model code is
NASA Astrophysics Data System (ADS)
Penenko, Vladimir; Tsvetova, Elena; Penenko, Alexey
2015-04-01
The proposed method is considered on an example of hydrothermodynamics and atmospheric chemistry models [1,2]. In the development of the existing methods for constructing numerical schemes possessing the properties of total approximation for operators of multiscale process models, we have developed a new variational technique, which uses the concept of adjoint integrating factors. The technique is as follows. First, a basic functional of the variational principle (the integral identity that unites the model equations, initial and boundary conditions) is transformed using Lagrange's identity and the second Green's formula. As a result, the action of the operators of main problem in the space of state functions is transferred to the adjoint operators defined in the space of sufficiently smooth adjoint functions. By the choice of adjoint functions the order of the derivatives becomes lower by one than those in the original equations. We obtain a set of new balance relationships that take into account the sources and boundary conditions. Next, we introduce the decomposition of the model domain into a set of finite volumes. For multi-dimensional non-stationary problems, this technique is applied in the framework of the variational principle and schemes of decomposition and splitting on the set of physical processes for each coordinate directions successively at each time step. For each direction within the finite volume, the analytical solutions of one-dimensional homogeneous adjoint equations are constructed. In this case, the solutions of adjoint equations serve as integrating factors. The results are the hybrid discrete-analytical schemes. They have the properties of stability, approximation and unconditional monotony for convection-diffusion operators. These schemes are discrete in time and analytic in the spatial variables. They are exact in case of piecewise-constant coefficients within the finite volume and along the coordinate lines of the grid area in each
Sensitivity analysis of numerically-simulated convective storms using direct and adjoint methods
Park, S.K.; Droegemeier, K.K.; Bischof, C.; Knauff, T.
1994-06-01
The goal of this project is to evaluate the sensitivity of numerically modeled convective storms to control parameters such as the initial conditions, boundary conditions, environment, and various physical and computational parameters. In other words, the authors seek the gradient of the solution vector with respect to specified parameters. One can use two approaches to accomplish this task. In the first or so-called brute force method, one uses a fully nonlinear model to generate a control forecast starting from a specified initial state. Then, a number of other forecasts are made in which chosen parameters (e.g., initial conditions) are systematically varied. The obvious drawback is that a large number of full model predictions are needed to examine the effects of only a single parameter. The authors describe herein an alternative, essentially automated method (ADIFOR, or Automatic DIfferentiation of FORtran) for obtaining the solution gradient that bypasses the adjoint altogether yet provides even more information about the gradient. (ADIFOR, like the adjoint technique, is constrained by the linearity assumption.) Applied to a 1-D moist cloud model, the authors assess the utility of ADIFOR relative to the brute force approach and evaluate the validity of the tangent linear approximation in the context of deep convection.
Reduction in CS: A (Mostly) Quantitative Analysis of Reductive Solutions to Algorithmic Problems
ERIC Educational Resources Information Center
Armoni, Michal
2009-01-01
Reduction is a problem-solving strategy, relevant to various areas of computer science, and strongly connected to abstraction: a reductive solution necessitates establishing a connection among problems that may seem totally disconnected at first sight, and abstracts the solution to the reduced-to problem by encapsulating it as a black box. The…
Fast Time and Space Parallel Algorithms for Solution of Parabolic Partial Differential Equations
NASA Technical Reports Server (NTRS)
Fijany, Amir
1993-01-01
In this paper, fast time- and Space -Parallel agorithms for solution of linear parabolic PDEs are developed. It is shown that the seemingly strictly serial iterations of the time-stepping procedure for solution of the problem can be completed decoupled.
Self-Adjoint Angular Flux Equation for Coupled Electron-Photon Transport
Liscum-Powell, J.L.; Lorence, L.J. Jr.; Morel, J.E.; Prinja, A.K.
1999-07-08
Recently, Morel and McGhee described an alternate second-order form of the transport equation called the self adjoint angular flux (SAAF) equation that has the angular flux as its unknown. The SAAF formulation has all the advantages of the traditional even- and odd-parity self-adjoint equations, with the added advantages that it yields the full angular flux when it is numerically solved, it is significantly easier to implement reflective and reflective-like boundary conditions, and in the appropriate form it can be solved in void regions. The SAAF equation has the disadvantage that the angular domain is the full unit sphere and, like the even- and odd- parity form, S{sub n} source iteration cannot be implemented using the standard sweeping algorithm. Also, problems arise in pure scattering media. Morel and McGhee demonstrated the efficacy of the SAAF formulation for neutral particle transport. Here we apply the SAAF formulation to coupled electron-photon transport problems using multigroup cross-sections from the CEPXS code and S{sub n} discretization.
Ocean acoustic tomography from different receiver geometries using the adjoint method.
Zhao, Xiaofeng; Wang, Dongxiao
2015-12-01
In this paper, an ocean acoustic tomography inversion using the adjoint method in a shallow water environment is presented. The propagation model used is an implicit Crank-Nicolson finite difference parabolic equation solver with a non-local boundary condition. Unlike previous matched-field processing works using the complex pressure fields as the observations, here, the observed signals are the transmission losses. Based on the code tests of the tangent linear model, the adjoint model, and the gradient, the optimization problem is solved by a gradient-based minimization algorithm. The inversions are performed in numerical simulations for two geometries: one in which hydrophones are sparsely distributed in the horizontal direction, and another in which the hydrophones are distributed vertically. The spacing in both cases is well beyond the half-wavelength threshold at which beamforming could be used. To deal with the ill-posedness of the inverse problem, a linear differential regularization operator of the sound-speed profile is used to smooth the inversion results. The L-curve criterion is adopted to select the regularization parameter, and the optimal value can be easily determined at the elbow of the logarithms of the residual norm of the measured-predicted fields and the norm of the penalty function. PMID:26723329
Park, H.; Densmore, J. D.; Wollaber, A. B.; Knoll, D. A.; Rauenzahn, R. M.
2013-07-01
We have developed a moment-based scale-bridging algorithm for thermal radiative transfer problems. The algorithm takes the form of well-known nonlinear-diffusion acceleration which utilizes a low-order (LO) continuum problem to accelerate the solution of a high-order (HO) kinetic problem. The coupled nonlinear equations that form the LO problem are efficiently solved using a preconditioned Jacobian-free Newton-Krylov method. This work demonstrates the applicability of the scale-bridging algorithm with a Monte Carlo HO solver and reports the computational efficiency of the algorithm in comparison to the well-known Fleck-Cummings algorithm. (authors)
The efficiency of geophysical adjoint codes generated by automatic differentiation tools
NASA Astrophysics Data System (ADS)
Vlasenko, A. V.; Köhl, A.; Stammer, D.
2016-02-01
The accuracy of numerical models that describe complex physical or chemical processes depends on the choice of model parameters. Estimating an optimal set of parameters by optimization algorithms requires knowledge of the sensitivity of the process of interest to model parameters. Typically the sensitivity computation involves differentiation of the model, which can be performed by applying algorithmic differentiation (AD) tools to the underlying numerical code. However, existing AD tools differ substantially in design, legibility and computational efficiency. In this study we show that, for geophysical data assimilation problems of varying complexity, the performance of adjoint codes generated by the existing AD tools (i) Open_AD, (ii) Tapenade, (iii) NAGWare and (iv) Transformation of Algorithms in Fortran (TAF) can be vastly different. Based on simple test problems, we evaluate the efficiency of each AD tool with respect to computational speed, accuracy of the adjoint, the efficiency of memory usage, and the capability of each AD tool to handle modern FORTRAN 90-95 elements such as structures and pointers, which are new elements that either combine groups of variables or provide aliases to memory addresses, respectively. We show that, while operator overloading tools are the only ones suitable for modern codes written in object-oriented programming languages, their computational efficiency lags behind source transformation by orders of magnitude, rendering the application of these modern tools to practical assimilation problems prohibitive. In contrast, the application of source transformation tools appears to be the most efficient choice, allowing handling even large geophysical data assimilation problems. However, they can only be applied to numerical models written in earlier generations of programming languages. Our study indicates that applying existing AD tools to realistic geophysical problems faces limitations that urgently need to be solved to allow the
Assimilating Remote Ammonia Observations with a Refined Aerosol Thermodynamics Adjoint"
Ammonia emissions parameters in North America can be refined in order to improve the evaluation of modeled concentrations against observations. Here, we seek to do so by developing and applying the GEOS-Chem adjoint nested over North America to conductassimilation of observations...
Adjoint Based Data Assimilation for an Ionospheric Model
NASA Astrophysics Data System (ADS)
Rosen, I. G.; Hajj, G. A.; Hajj, G. A.; Pi, X.; Pi, X.; Wang, C.; Wilson, B. D.
2001-05-01
The success of ionospheric modeling depends primarily on accurate knowledge of the forces (drivers) which enter into the collisional plasma hydrodynamic equations for the ionosphere and control the ionization as well as other dynamical and chemical processes. These include solar EUV and UV radiation, magnetospheric electric fields, particle precipitation, dynamo electric fields, thermospheric winds, neutral densities, and temperature. The determination of these model parameters from observational data is known as data assimilation. The data assimilation problem is formulated as a problem of minimizing a nonlinear functional, J (typically least squares) under a system of constraints consisting primarily of the underlying model equations. The performance index, J, can, in principle, be minimized using standard techniques such as the Newton's steepest decent method. There are however major technical challenges in practice. Since J is highly nonlinear and each evaluation of the functional requires the integration of the ionospheric model equations, computing the gradient vector of J with respect to the unknown parameters is time consuming. This problem is solved by use of the adjoint method. The ionospheric model used in this effort is for mid- and low-latitudes and consists of solving the continuity and momentum partial differential equations in four dimensional (three spatial dimensions and time) to compute the O+ density in the ionosphere and plasmasphere. We have developed codes for solving the forward model on a fixed grid. This makes it relatively straight forward to apply the adjoint method for computing gradients when doing nonlinear least squares based data assimilation. Because of the significant cost (in computational effort and CPU time) involved in performing a forward integration of the underlying 3-D model at any reasonable grid resolution, the use of the adjoint method for computing the gradients is indispensable. The adjoint method provides an elegant
NASA Astrophysics Data System (ADS)
García, Hermes A.; Guerrero-Bolaño, Francisco J.; Obregón-Neira, Nelson
2010-05-01
Due to both mathematical tractability and efficiency on computational resources, it is very common to find in the realm of numerical modeling in hydro-engineering that regular linearization techniques have been applied to nonlinear partial differential equations properly obtained in environmental flow studies. Sometimes this simplification is also made along with omission of nonlinear terms involved in such equations which in turn diminishes the performance of any implemented approach. This is the case for example, for contaminant transport modeling in streams. Nowadays, a traditional and one of the most common used water quality model such as QUAL2k, preserves its original algorithm, which omits nonlinear terms through linearization techniques, in spite of the continuous algorithmic development and computer power enhancement. For that reason, the main objective of this research was to generate a flexible tool for non-linear water quality modeling. The solution implemented here was based on two genetic algorithms, used in a nested way in order to find two different types of solutions sets: the first set is composed by the concentrations of the physical-chemical variables used in the modeling approach (16 variables), which satisfies the non-linear equation system. The second set, is the typical solution of the inverse problem, the parameters and constants values for the model when it is applied to a particular stream. From a total of sixteen (16) variables, thirteen (13) was modeled by using non-linear coupled equation systems and three (3) was modeled in an independent way. The model used here had a requirement of fifty (50) parameters. The nested genetic algorithm used for the numerical solution of a non-linear equation system proved to serve as a flexible tool to handle with the intrinsic non-linearity that emerges from the interactions occurring between multiple variables involved in water quality studies. However because there is a strong data limitation in
Adaptive mesh refinement and adjoint methods in geophysics simulations
NASA Astrophysics Data System (ADS)
Burstedde, Carsten
2013-04-01
required by human intervention and analysis. Specifying an objective functional that quantifies the misfit between the simulation outcome and known constraints and then minimizing it through numerical optimization can serve as an automated technique for parameter identification. As suggested by the similarity in formulation, the numerical algorithm is closely related to the one used for goal-oriented error estimation. One common point is that the so-called adjoint equation needs to be solved numerically. We will outline the derivation and implementation of these methods and discuss some of their pros and cons, supported by numerical results.
NASA Astrophysics Data System (ADS)
Hagedoorn, J. M.; Martinec, Z.
2012-12-01
Recent models of the Earth's geomagnetic field at the core-mantle boundary (CMB) are based on satellite measurements and/or observatory data, which are mostly harmonically downward continued to the CMB. One aim of the upcoming satellite mission Swarm is to determine the three-dimensional distribution of electric conductivity of the Earth's mantle. On this background, we developed an adjoint sensitivity downward continuation approach that is capable to consider three-dimensional electric conductivity distributions. Martinec (Geophys. J. Int., 136, 1999) developed a time-domain spectral-finite element approach for the forward modelling of vector electromagnetic induction data as measured on ground-based magnetic observatory or by satellites. We design a new method to compute the sensitivity of the magnetic induction data to a magnetic field prescribed at the core-mantle boundary, which we term the adjoint sensitivity method. The forward and adjoint initial boundary-value problems, both solved in the time domain, are identical, except for the specification of prescribed boundary conditions. The respective boundary-value data are the measured X magnetic component for the forward method and the difference between the measured and predicted Z magnetic component for the adjoint method. The squares of the differences in Z magnetic component summed up over the time of observation and all spatial positions of observations determine the misfit. Then the sensitivities of observed data, i.e. the partial derivatives of the misfit with respect to the parameters characterizing the magnetic field at the core-mantle boundary, are obtained by the surface integral over the core-mantle boundary of the product of the adjoint solution multiplied by the time-dependent functions describing the time variability of magnetic field at the core-mantle boundary, and integrated over the time of observation. The time variability of boundary data is represented in terms of locally supported B
NASA Astrophysics Data System (ADS)
Ouwerling, G. J. L.
1989-12-01
This paper gives a concise overview of some existing methods for the solution of the nonlinear Poisson equation in semiconductors. A method for the solution of this equation was recently proposed in this journal by I. D. Mayergoyz [J. Appl. Phys. 59, 195 (1986)]. Soon afterwards, an improved version was described by W. Keller [J. Appl. Phys. 61, 5189 (1987)]. Both methods are classified within the perspective of the existing methods. Moreover, Keller's method is further improved by the introduction of scaled variables and by using red-black ordening to allow for overrelaxation. All advantages of the two methods are maintained. An illustrative example shows an improvement in solution speed of at least a factor of 5.6.
A local anisotropic adaptive algorithm for the solution of low-Mach transient combustion problems
NASA Astrophysics Data System (ADS)
Carpio, Jaime; Prieto, Juan Luis; Vera, Marcos
2016-02-01
A novel numerical algorithm for the simulation of transient combustion problems at low Mach and moderately high Reynolds numbers is presented. These problems are often characterized by the existence of a large disparity of length and time scales, resulting in the development of directional flow features, such as slender jets, boundary layers, mixing layers, or flame fronts. This makes local anisotropic adaptive techniques quite advantageous computationally. In this work we propose a local anisotropic refinement algorithm using, for the spatial discretization, unstructured triangular elements in a finite element framework. For the time integration, the problem is formulated in the context of semi-Lagrangian schemes, introducing the semi-Lagrange-Galerkin (SLG) technique as a better alternative to the classical semi-Lagrangian (SL) interpolation. The good performance of the numerical algorithm is illustrated by solving a canonical laminar combustion problem: the flame/vortex interaction. First, a premixed methane-air flame/vortex interaction with simplified transport and chemistry description (Test I) is considered. Results are found to be in excellent agreement with those in the literature, proving the superior performance of the SLG scheme when compared with the classical SL technique, and the advantage of using anisotropic adaptation instead of uniform meshes or isotropic mesh refinement. As a more realistic example, we then conduct simulations of non-premixed hydrogen-air flame/vortex interactions (Test II) using a more complex combustion model which involves state-of-the-art transport and chemical kinetics. In addition to the analysis of the numerical features, this second example allows us to perform a satisfactory comparison with experimental visualizations taken from the literature.
Solid oxide fuel cell simulation and design optimization with numerical adjoint techniques
NASA Astrophysics Data System (ADS)
Elliott, Louie C.
This dissertation reports on the application of numerical optimization techniques as applied to fuel cell simulation and design. Due to the "multi-physics" inherent in a fuel cell, which results in a highly coupled and non-linear behavior, an experimental program to analyze and improve the performance of fuel cells is extremely difficult. This program applies new optimization techniques with computational methods from the field of aerospace engineering to the fuel cell design problem. After an overview of fuel cell history, importance, and classification, a mathematical model of solid oxide fuel cells (SOFC) is presented. The governing equations are discretized and solved with computational fluid dynamics (CFD) techniques including unstructured meshes, non-linear solution methods, numerical derivatives with complex variables, and sensitivity analysis with adjoint methods. Following the validation of the fuel cell model in 2-D and 3-D, the results of the sensitivity analysis are presented. The sensitivity derivative for a cost function with respect to a design variable is found with three increasingly sophisticated techniques: finite difference, direct differentiation, and adjoint. A design cycle is performed using a simple optimization method to improve the value of the implemented cost function. The results from this program could improve fuel cell performance and lessen the world's dependence on fossil fuels.
Adjoint Tomography of Taiwan Region: From Travel-Time Toward Waveform Inversion
NASA Astrophysics Data System (ADS)
Huang, H. H.; Lee, S. J.; Tromp, J.
2014-12-01
The complicated tectonic environment such as Taiwan region can modulate the seismic waveform severely and hamper the discrimination and the utilization of later phases. Restricted to the use of only first arrivals of P- and S-wave, the travel-time tomographic models of Taiwan can simulate the seismic waveform barely to a frequency of 0.2 Hz to date. While it has been sufficient for long-period studies, e.g. source inversion, this frequency band is still far from the applications to the community and high-resolution studies. To achieve a higher-frequency simulation, more data and the considerations of off-path and finite-frequency effects are necessary. Based on the spectral-element and the adjoint method recently developed, we prepared 94 MW 3.5-6.0 earthquakes with well-defined location and focal mechanism solutions from Real-Time Moment Tensor Monitoring System (RMT), and preformed an iterative gradient-based inversion employing waveform modeling and finite-frequency measurements of adjoint method. By which the 3-D sensitivity kernels are taken into account realistically and the full waveform information are naturally sought, without a need of any phase pick. A preliminary model m003 using 10-50 sec data was demonstrated and compared with previous travel-time models. The primary difference appears in the mountainous area, where the previous travel-time model may underestimate the S-wave speed in the upper crust, but overestimates in the lower crust.
NASA Astrophysics Data System (ADS)
Yu, Jia; Ji, Lucheng; Li, Weiwei; Yi, Weilin
2016-06-01
Adjoint method is an important tool for design refinement of multistage compressors. However, the radial static pressure distribution deviates during the optimization procedure and deteriorates the overall performance, producing final designs that are not well suited for realistic engineering applications. In previous development work on multistage turbomachinery blade optimization using adjoint method and thin shear-layer N-S equations, the entropy production is selected as the objective function with given mass flow rate and total pressure ratio as imposed constraints. The radial static pressure distribution at the interfaces between rows is introduced as a new constraint in the present paper. The approach is applied to the redesign of a five-stage axial compressor, and the results obtained with and without the constraint on the radial static pressure distribution at the interfaces between rows are discussed in detail. The results show that the redesign without the radial static pressure distribution constraint (RSPDC) gives an optimal solution that shows deviations on radial static pressure distribution, especially at rotor exit tip region. On the other hand, the redesign with the RSPDC successfully keeps the radial static pressure distribution at the interfaces between rows and make sure that the optimization results are applicable in a practical engineering design.
Aerodynamic Shape Optimization of Complex Aircraft Configurations via an Adjoint Formulation
NASA Technical Reports Server (NTRS)
Reuther, James; Jameson, Antony; Farmer, James; Martinelli, Luigi; Saunders, David
1996-01-01
This work describes the implementation of optimization techniques based on control theory for complex aircraft configurations. Here control theory is employed to derive the adjoint differential equations, the solution of which allows for a drastic reduction in computational costs over previous design methods (13, 12, 43, 38). In our earlier studies (19, 20, 22, 23, 39, 25, 40, 41, 42) it was shown that this method could be used to devise effective optimization procedures for airfoils, wings and wing-bodies subject to either analytic or arbitrary meshes. Design formulations for both potential flows and flows governed by the Euler equations have been demonstrated, showing that such methods can be devised for various governing equations (39, 25). In our most recent works (40, 42) the method was extended to treat wing-body configurations with a large number of mesh points, verifying that significant computational savings can be gained for practical design problems. In this paper the method is extended for the Euler equations to treat complete aircraft configurations via a new multiblock implementation. New elements include a multiblock-multigrid flow solver, a multiblock-multigrid adjoint solver, and a multiblock mesh perturbation scheme. Two design examples are presented in which the new method is used for the wing redesign of a transonic business jet.
2014-01-01
Background The 3D chromatogram generated by High Performance Liquid Chromatography-Diode Array Detector (HPLC-DAD) has been researched widely in the field of herbal medicine, grape wine, agriculture, petroleum and so on. Currently, most of the methods used for separating a 3D chromatogram need to know the compounds' number in advance, which could be impossible especially when the compounds are complex or white noise exist. New method which extracts compounds from 3D chromatogram directly is needed. Methods In this paper, a new separation model named parallel Independent Component Analysis constrained by Reference Curve (pICARC) was proposed to transform the separation problem to a multi-parameter optimization issue. It was not necessary to know the number of compounds in the optimization. In order to find all the solutions, an algorithm named multi-areas Genetic Algorithm (mGA) was proposed, where multiple areas of candidate solutions were constructed according to the fitness and distances among the chromosomes. Results Simulations and experiments on a real life HPLC-DAD data set were used to demonstrate our method and its effectiveness. Through simulations, it can be seen that our method can separate 3D chromatogram to chromatogram peaks and spectra successfully even when they severely overlapped. It is also shown by the experiments that our method is effective to solve real HPLC-DAD data set. Conclusions Our method can separate 3D chromatogram successfully without knowing the compounds' number in advance, which is fast and effective. PMID:25474487
Debbarma, Sanjoy; Saikia, Lalit Chandra; Sinha, Nidul
2014-03-01
Present work focused on automatic generation control (AGC) of a three unequal area thermal systems considering reheat turbines and appropriate generation rate constraints (GRC). A fractional order (FO) controller named as I(λ)D(µ) controller based on crone approximation is proposed for the first time as an appropriate technique to solve the multi-area AGC problem in power systems. A recently developed metaheuristic algorithm known as firefly algorithm (FA) is used for the simultaneous optimization of the gains and other parameters such as order of integrator (λ) and differentiator (μ) of I(λ)D(µ) controller and governor speed regulation parameters (R). The dynamic responses corresponding to optimized I(λ)D(µ) controller gains, λ, μ, and R are compared with that of classical integer order (IO) controllers such as I, PI and PID controllers. Simulation results show that the proposed I(λ)D(µ) controller provides more improved dynamic responses and outperforms the IO based classical controllers. Further, sensitivity analysis confirms the robustness of the so optimized I(λ)D(µ) controller to wide changes in system loading conditions and size and position of SLP. Proposed controller is also found to have performed well as compared to IO based controllers when SLP takes place simultaneously in any two areas or all the areas. Robustness of the proposed I(λ)D(µ) controller is also tested against system parameter variations. PMID:24139308
NASA Astrophysics Data System (ADS)
Kopacz, Monika; Jacob, Daniel J.; Henze, Daven K.; Heald, Colette L.; Streets, David G.; Zhang, Qiang
2008-04-01
We apply the adjoint of an atmospheric chemical transport model (GEOS-Chem CTM) to constrain Asian sources of carbon monoxide (CO) with 2° × 2.5° spatial resolution using Measurement of Pollution in the Troposphere (MOPITT) satellite observations of CO columns in February-April 2001. Results are compared to the more common analytical method for solving the same Bayesian inverse problem and applied to the same data set. The analytical method is more exact but because of computational limitations it can only constrain emissions over coarse regions. We find that the correction factors to the a priori CO emission inventory from the adjoint inversion are consistent with those of the analytical inversion when averaged over the large regions of the latter. Unlike the analytical solution, the adjoint correction factors are not subject to compensating errors between adjacent regions (error anticorrelation). The adjoint solution also reveals fine-scale variability that the analytical inversion cannot resolve. For example, India shows both large emissions underestimates in the densely populated Ganges Valley and large overestimates in the eastern part of the country where springtime emissions are dominated by biomass burning. Correction factors to Chinese emissions are largest in central and eastern China, consistent with a recent bottom-up inventory though there are disagreements in the fine structure. Correction factors for biomass burning are consistent with a recent bottom-up inventory based on MODIS satellite fire data.
Towards adjoint-based inversion for rheological parameters in nonlinear viscous mantle flow
NASA Astrophysics Data System (ADS)
Worthen, Jennifer; Stadler, Georg; Petra, Noemi; Gurnis, Michael; Ghattas, Omar
2014-09-01
We address the problem of inferring mantle rheological parameter fields from surface velocity observations and instantaneous nonlinear mantle flow models. We formulate this inverse problem as an infinite-dimensional nonlinear least squares optimization problem governed by nonlinear Stokes equations. We provide expressions for the gradient of the cost functional of this optimization problem with respect to two spatially-varying rheological parameter fields: the viscosity prefactor and the exponent of the second invariant of the strain rate tensor. Adjoint (linearized) Stokes equations, which are characterized by a 4th order anisotropic viscosity tensor, facilitates efficient computation of the gradient. A quasi-Newton method for the solution of this optimization problem is presented, which requires the repeated solution of both nonlinear forward Stokes and linearized adjoint Stokes equations. For the solution of the nonlinear Stokes equations, we find that Newton’s method is significantly more efficient than a Picard fixed point method. Spectral analysis of the inverse operator given by the Hessian of the optimization problem reveals that the numerical eigenvalues collapse rapidly to zero, suggesting a high degree of ill-posedness of the inverse problem. To overcome this ill-posedness, we employ Tikhonov regularization (favoring smooth parameter fields) or total variation (TV) regularization (favoring piecewise-smooth parameter fields). Solution of two- and three-dimensional finite element-based model inverse problems show that a constant parameter in the constitutive law can be recovered well from surface velocity observations. Inverting for a spatially-varying parameter field leads to its reasonable recovery, in particular close to the surface. When inferring two spatially varying parameter fields, only an effective viscosity field and the total viscous dissipation are recoverable. Finally, a model of a subducting plate shows that a localized weak zone at the
Towards Multi-resolution Adjoint Tomography of the European Crust and Upper Mantle
NASA Astrophysics Data System (ADS)
Basini, P.; Nissen-Meyer, T.; Boschi, L.; Schenk, O.; Verbeke, J.; Hanasoge, S.; Giardini, D.
2010-12-01
Thanks to continuously improved instrument coverage, and the growth of high-performance computational infrastructure, it is now possible to enhance the resolution at which seismologists image the Earth's interior. While most algorithms in global seismic tomography today are grounded on the ray-theory approximation, however, resolution and model complexity can effectively be enhanced only through the application of more advanced techniques accounting for the many complexities of the partial derivatives relating seismic data and Earth structure. These include full-wave forward modelling methods and adjoint algorithms, which together set a framework for iterative, nonlinear inversion upon complex 3D structures. We take advantage of these methodological improvements using a newly developed, flexible spectral-element method (SPECFEM3D) with embedded adjoint capabilities to devise new tomographic models of the European crust and upper mantle. We chose a two-scale strategy, in which we use global surface wave data to first constrain the large-scale structures, and simultaneously invert for high-resolution, regional structures based on measurements of ambient noise in central and southern Europe. By its very nature, and as a result of the dense station coverage over the continent, the ambient-noise method affords us a particularly uniform seismic coverage. To define surface-wave sensitivity kernels, we construct a flexible, global mesh of the upper mantle only (i.e., a spherical shell) honoring all global discontinuities, and include 3D starting models down to periods of 30 seconds. The noise data are cross-correlated to obtain station-to-station Green's functions. We will present examples of sensitivity kernels computed for these noise-based Green's functions and discuss the data-specific validity of the underlying assumptions to extract Green's functions. The local setup, which is constructed using the same software as in the global case, needs to honor internal and
Development of Web-Based Menu Planning Support System and its Solution Using Genetic Algorithm
NASA Astrophysics Data System (ADS)
Kashima, Tomoko; Matsumoto, Shimpei; Ishii, Hiroaki
2009-10-01
Recently lifestyle-related diseases have become an object of public concern, while at the same time people are being more health conscious. As an essential factor for causing the lifestyle-related diseases, we assume that the knowledge circulation on dietary habits is still insufficient. This paper focuses on everyday meals close to our life and proposes a well-balanced menu planning system as a preventive measure of lifestyle-related diseases. The system is developed by using a Web-based frontend and it provides multi-user services and menu information sharing capabilities like social networking services (SNS). The system is implemented on a Web server running Apache (HTTP server software), MySQL (database management system), and PHP (scripting language for dynamic Web pages). For the menu planning, a genetic algorithm is applied by understanding this problem as multidimensional 0-1 integer programming.
Chacón, P; Morán, F; Díaz, J F; Pantos, E; Andreu, J M
1998-01-01
Small-angle x-ray solution scattering (SAXS) is analyzed with a new method to retrieve convergent model structures that fit the scattering profiles. An arbitrary hexagonal packing of several hundred beads containing the problem object is defined. Instead of attempting to compute the Debye formula for all of the possible mass distributions, a genetic algorithm is employed that efficiently searches the configurational space and evolves best-fit bead models. Models from different runs of the algorithm have similar or identical structures. The modeling resolution is increased by reducing the bead radius together with the search space in successive cycles of refinement. The method has been tested with protein SAXS (0.001 < S < 0.06 A(-1)) calculated from x-ray crystal structures, adding noise to the profiles. The models obtained closely approach the volumes and radii of gyration of the known structures, and faithfully reproduce the dimensions and shape of each of them. This includes finding the active site cavity of lysozyme, the bilobed structure of gamma-crystallin, two domains connected by a stalk in betab2-crystallin, and the horseshoe shape of pancreatic ribonuclease inhibitor. The low-resolution solution structure of lysozyme has been directly modeled from its experimental SAXS profile (0.003 < S < 0.03 A(-1)). The model describes lysozyme size and shape to the resolution of the measurement. The method may be applied to other proteins, to the analysis of domain movements, to the comparison of solution and crystal structures, as well as to large macromolecular assemblies. PMID:9635731
NASA Astrophysics Data System (ADS)
Holdaway, D. R.; Errico, R.
2011-12-01
Inherent in the minimization process in the 4D-Var data assimilation system is the need for the model's adjoint. It is straightforward to obtain the exact adjoint by linearizing the code in a line by line sense; however it only provides an accurate overall representation of the physical processes if the model behaviour is linear. Moist processes in the atmosphere, and thus the models that represent them, are intrinsically highly non-linear and can contain discrete switches. The adjoint that is required in the data assimilation system needs to provide an accurate representation of the physical behaviour for perturbation sizes of the order of the analysis error, so an exact adjoint of the moist physics model is likely to be inaccurate. Instead a non-exact adjoint model, which is accurate for large enough perturbations, must be developed. The constraint on the development is that the simplified adjoint be consistent with the actual trajectory of the model. Previous attempts to include the moist physics in the 4D-Var have emphasized the need for redevelopment of the actual moist scheme to a simpler version. These schemes are designed to be linear in the limit of realistic perturbation size but also capture the essence of the physical behaviour, making the adjoint version of the scheme suitable for use in the 4D-Var. A downside to this approach is that it can result in an over simplification of the physics and represent a larger departure from the true model trajectory than necessary. The adjoint is just the transpose of the tangent linear model, which is the differential of the model operator. This differential of the operator can be constructed from Jacobian matrices. Examining the structures of the Jacobians as perturbations of varying size are added to the state vector can help determine whether the adjoint model - be it of actual or simplified physics - will be suitable for use in the assimilation algorithm. If Jacobian structures change considerably when the
Fischer, P.F.
1996-12-31
Efficient solution of the Navier-Stokes equations in complex domains is dependent upon the availability of fast solvers for sparse linear systems. For unsteady incompressible flows, the pressure operator is the leading contributor to stiffness, as the characteristic propagation speed is infinite. In the context of operator splitting formulations, it is the pressure solve which is the most computationally challenging, despite its elliptic origins. We seek to improve existing spectral element iterative methods for the pressure solve in order to overcome the slow convergence frequently observed in the presence of highly refined grids or high-aspect ratio elements.
Accurate adjoint design sensitivities for nano metal optics.
Hansen, Paul; Hesselink, Lambertus
2015-09-01
We present a method for obtaining accurate numerical design sensitivities for metal-optical nanostructures. Adjoint design sensitivity analysis, long used in fluid mechanics and mechanical engineering for both optimization and structural analysis, is beginning to be used for nano-optics design, but it fails for sharp-cornered metal structures because the numerical error in electromagnetic simulations of metal structures is highest at sharp corners. These locations feature strong field enhancement and contribute strongly to design sensitivities. By using high-accuracy FEM calculations and rounding sharp features to a finite radius of curvature we obtain highly-accurate design sensitivities for 3D metal devices. To provide a bridge to the existing literature on adjoint methods in other fields, we derive the sensitivity equations for Maxwell's equations in the PDE framework widely used in fluid mechanics. PMID:26368483
Examination of Observation Impacts derived from OSEs and Adjoint Models
NASA Technical Reports Server (NTRS)
Gelaro, Ronald
2008-01-01
With the adjoint of a data assimilation system, the impact of any or all assimilated observations on measures of forecast skill can be estimated accurately and efficiently. The approach allows aggregation of results in terms of individual data types, channels or locations, all computed simultaneously. In this study, adjoint-based estimates of observation impact are compared with results from standard observing system experiments (OSEs) in the NASA Goddard Earth Observing System Model, Version 5 (GEOS-5) GEOS-5 system. The two approaches are shown to provide unique, but complimentary, information. Used together, they reveal both redundancies and dependencies between observing system impacts as observations are added or removed. Understanding these dependencies poses a major challenge for optimizing the use of the current observational network and defining requirements for future observing systems.
Three-Dimensional Turbulent RANS Adjoint-Based Error Correction
NASA Technical Reports Server (NTRS)
Park, Michael A.
2003-01-01
Engineering problems commonly require functional outputs of computational fluid dynamics (CFD) simulations with specified accuracy. These simulations are performed with limited computational resources. Computable error estimates offer the possibility of quantifying accuracy on a given mesh and predicting a fine grid functional on a coarser mesh. Such an estimate can be computed by solving the flow equations and the associated adjoint problem for the functional of interest. An adjoint-based error correction procedure is demonstrated for transonic inviscid and subsonic laminar and turbulent flow. A mesh adaptation procedure is formulated to target uncertainty in the corrected functional and terminate when error remaining in the calculation is less than a user-specified error tolerance. This adaptation scheme is shown to yield anisotropic meshes with corrected functionals that are more accurate for a given number of grid points then isotropic adapted and uniformly refined grids.
On improving storm surge forecasting using an adjoint optimal technique
NASA Astrophysics Data System (ADS)
Li, Yineng; Peng, Shiqiu; Yan, Jing; Xie, Lian
2013-12-01
A three-dimensional ocean model and its adjoint model are used to simultaneously optimize the initial conditions (IC) and the wind stress drag coefficient (Cd) for improving storm surge forecasting. To demonstrate the effect of this proposed method, a number of identical twin experiments (ITEs) with a prescription of different error sources and two real data assimilation experiments are performed. Results from both the idealized and real data assimilation experiments show that adjusting IC and Cd simultaneously can achieve much more improvements in storm surge forecasting than adjusting IC or Cd only. A diagnosis on the dynamical balance indicates that adjusting IC only may introduce unrealistic oscillations out of the assimilation window, which can be suppressed by the adjustment of the wind stress when simultaneously adjusting IC and Cd. Therefore, it is recommended to simultaneously adjust IC and Cd to improve storm surge forecasting using an adjoint technique.
Hep, J.; Konecna, A.; Krysl, V.; Smutny, V.
2011-07-01
This paper describes the application of effective source in forward calculations and the adjoint method to the solution of fast neutron fluence and activation detector activities in the reactor pressure vessel (RPV) and RPV cavity of a VVER-440 reactor. Its objective is the demonstration of both methods on a practical task. The effective source method applies the Boltzmann transport operator to time integrated source data in order to obtain neutron fluence and detector activities. By weighting the source data by time dependent decay of the detector activity, the result of the calculation is the detector activity. Alternatively, if the weighting is uniform with respect to time, the result is the fluence. The approach works because of the inherent linearity of radiation transport in non-multiplying time-invariant media. Integrated in this way, the source data are referred to as the effective source. The effective source in the forward calculations method thereby enables the analyst to replace numerous intensive transport calculations with a single transport calculation in which the time dependence and magnitude of the source are correctly represented. In this work, the effective source method has been expanded slightly in the following way: neutron source data were performed with few group method calculation using the active core calculation code MOBY-DICK. The follow-up neutron transport calculation was performed using the neutron transport code TORT to perform multigroup calculations. For comparison, an alternative method of calculation has been used based upon adjoint functions of the Boltzmann transport equation. Calculation of the three-dimensional (3-D) adjoint function for each required computational outcome has been obtained using the deterministic code TORT and the cross section library BGL440. Adjoint functions appropriate to the required fast neutron flux density and neutron reaction rates have been calculated for several significant points within the RPV
Benchmarking algorithms for the solution of Collisional Radiative Model (CRM) equations.
NASA Astrophysics Data System (ADS)
Klapisch, Marcel; Busquet, Michel
2007-11-01
Elements used in ICF target designs can have many charge states in the same plasma conditions, each charge state having numerous energy levels. When LTE conditions are not met, one has to solve CRM equations for the populations of energy levels, which are necessary for opacities/emissivities, Z* etc. In case of sparse spectra, or when configuration interaction is important (open d or f shells), statistical methods[1] are insufficient. For these cases one must resort to a detailed level CRM rate generator[2]. The equations to be solved may involve tens of thousands of levels. The system is by nature ill conditioned. We show that some classical methods do not converge. Improvements of the latter will be compared with new algorithms[3] with respect to performance, robustness, and accuracy. [1] A Bar-Shalom, J Oreg, and M Klapisch, J. Q. S. R. T.,65, 43 (2000). [2] M Klapisch, M Busquet and A. Bar-Shalom, Proceedings of APIP'07, AIP series (to be published). [3] M Klapisch and M Busquet, High Ener. Density Phys. 3,143 (2007)
Johnson, S A; Zhou, Y; Tracy, M K; Berggren, M J; Stenger, F
1984-01-01
olving the inverse scattering problem for the Helmholtz wave equation without employing the Born or Rytov approximations is a challenging problem, but some slow iterative methods have been proposed. One such method suggested by us is based on solving systems of nonlinear algebraic equations that are derived by applying the method of moments to a sinc basis function expansion of the fields and scattering potential. In the past, we have solved these equations for a 2-D object of n by n pixels in a time proportional to n5. In the present paper, we demonstrate a new method based on FFT convolution and the concept of backprojection which solves these equations in time proportional to n3 X log(n). Several numerical examples are given for images up to 7 by 7 pixels in size. Analogous algorithms to solve the Riccati wave equation in n3 X log(n) time are also suggested, but not verified. A method is suggested for interpolating measurements from one detector geometry to a new perturbed detector geometry whose measurement points fall on a FFT accessible, rectangular grid and thereby render many detector geometrics compatible for use by our fast methods. PMID:6540908
Seismic Window Selection and Misfit Measurements for Global Adjoint Tomography
NASA Astrophysics Data System (ADS)
Lei, W.; Bozdag, E.; Lefebvre, M.; Podhorszki, N.; Smith, J. A.; Tromp, J.
2013-12-01
Global Adjoint Tomography requires fast parallel processing of large datasets. After obtaing the preprocessed observed and synthetic seismograms, we use the open source software packages FLEXWIN (Maggi et al. 2007) to select time windows and MEASURE_ADJ to make measurements. These measurements define adjoint sources for data assimilation. Previous versions of these tools work on a pair of SAC files---observed and synthetic seismic data for the same component and station, and loop over all seismic records associated with one earthquake. Given the large number of stations and earthquakes, the frequent read and write operations create severe I/O bottlenecks on modern computing platforms. We present new versions of these tools utilizing a new seismic data format, namely the Adaptive Seismic Data Format(ASDF). This new format shows superior scalability for applications on high-performance computers and accommodates various types of data, including earthquake, industry and seismic interferometry datasets. ASDF also provides user-friendly APIs, which can be easily integrated into the adjoint tomography workflow and combined with other data processing tools. In addition to solving the I/O bottleneck, we are making several improvements to these tools. For example, FLEXWIN is tuned to select windows for different types of earthquakes. To capture their distinct features, we categorize earthquakes by their depths and frequency bands. Moreover, instead of only picking phases between the first P arrival and the surface-wave arrivals, our aim is to select and assimilate many other later prominent phases in adjoint tomography. For example, in the body-wave band (17 s - 60 s), we include SKS, sSKS and their multiple, while in the surface-wave band (60 s - 120 s) we incorporate major-arc surface waves.
A comparison of adjoint and data-centric verification techniques.
Wildey, Timothy Michael; Cyr, Eric Christopher; Shadid, John Nicolas; Pawlowski, Roger Patrick; Smith, Thomas Michael
2013-03-01
This document summarizes the results from a level 3 milestone study within the CASL VUQ effort. We compare the adjoint-based a posteriori error estimation approach with a recent variant of a data-centric verification technique. We provide a brief overview of each technique and then we discuss their relative advantages and disadvantages. We use Drekar::CFD to produce numerical results for steady-state Navier Stokes and SARANS approximations. 3
Monopole condensation in two-flavor adjoint QCD
Cossu, Guido; D'Elia, Massimo; Di Giacomo, Adriano; Lacagnina, Giuseppe; Pica, Claudio
2008-04-01
In QCD with adjoint fermions, the deconfining transition takes place at a lower temperature than the chiral transition. We study the two transitions by use of the Polyakov loop, the monopole order parameter, and the chiral condensate. The deconfining transition is first order, the chiral is a crossover. The order parameters for confinement are not affected by the chiral transition. We conclude that the degrees of freedom relevant to confinement are different from those describing chiral symmetry.
Spectral monodromy of non-self-adjoint operators
NASA Astrophysics Data System (ADS)
Phan, Quang Sang
2014-01-01
In the present paper, we build a combinatorial invariant, called the "spectral monodromy" from the spectrum of a single (non-self-adjoint) h-pseudodifferential operator with two degrees of freedom in the semi-classical limit. Our inspiration comes from the quantum monodromy defined for the joint spectrum of an integrable system of n commuting self-adjoint h-pseudodifferential operators, given by S. Vu Ngoc ["Quantum monodromy in integrable systems," Commun. Math. Phys. 203(2), 465-479 (1999)]. The first simple case that we treat in this work is a normal operator. In this case, the discrete spectrum can be identified with the joint spectrum of an integrable quantum system. The second more complex case we propose is a small perturbation of a self-adjoint operator with a classical integrability property. We show that the discrete spectrum (in a small band around the real axis) also has a combinatorial monodromy. The main difficulty in this case is that we do not know the description of the spectrum everywhere, but only in a Cantor type set. In addition, we also show that the corresponding monodromy can be identified with the classical monodromy, defined by J. Duistermaat ["On global action-angle coordinates," Commun. Pure Appl. Math. 33(6), 687-706 (1980)].
Adjoint-based sensitivity analysis for reactor-safety applications
Parks, C.V.
1985-01-01
The application and usefulness of an adjoint-based methodology for performing sensitivity analysis on reactor safety computer codes is investigated. The adjoint-based methodology, referred to as differential sensitivity theory (DST), provides first-order derivatives of the calculated quantities of interest (responses) with respect to the input parameters. The basic theoretical development of DST is presented along with the needed general extensions for consideration of model discontinuities and a variety of useful response definitions. A simple analytic problem is used to highlight the general DST procedures. Finally, DST procedures presented in this work are applied to two highly nonlinear reactor accident analysis codes: (1) FASTGAS, a relatively small code for analysis of loss-of-decay-heat-removal accident in a gas-cooled fast reactor, and (2) an existing code called VENUS-II which is typically employed for analyzing the core disassembly phase of a hypothetical fast reactor accident. The two codes are different both in terms of complexity and in terms of the facets of DST which can be illustrated. Sensitivity results from the adjoint codes ADJGAS and VENUS-ADJ are verified with direct recalculations using perturbed input parameters. The effectiveness of the DST results for parameter ranking, prediction of response changes, and uncertainty analysis are illustrated. The conclusion drawn from this study is that DST is a viable, cost-effective methodology for accurate sensitivity analysis.
Self-adjoint time operators and invariant subspaces
NASA Astrophysics Data System (ADS)
Gómez, Fernando
2008-02-01
The question of existence of self-adjoint time operators for unitary evolutions in classical and quantum mechanics is revisited on the basis of Halmos-Helson theory of invariant subspaces, Sz.-Nagy-Foiaş dilation theory and Misra-Prigogine-Courbage theory of irreversibility. It is shown that the existence of self-adjoint time operators is equivalent to the intertwining property of the evolution plus the existence of simply invariant subspaces or rigid operator-valued functions for its Sz.-Nagy-Foiaş functional model. Similar equivalent conditions are given in terms of intrinsic randomness in the context of statistical mechanics. The rest of the contents are mainly a unifying review of the subject scattered throughout an unconnected literature. A well-known extensive set of equivalent conditions is derived from the above results; such conditions are written in terms of Schrrdinger couples, the Weyl commutation relation, incoming and outgoing subspaces, innovation processes, Lax-Phillips scattering, translation and spectral representations, and spectral properties. Also the natural procedure dealing with symmetric time operators in standard quantum mechanics involving their self-adjoint extensions is illustrated by considering the quantum Aharonov-Bohm time-of-arrival operator.
Unsteady Adjoint Approach for Design Optimization of Flapping Airfoils
NASA Technical Reports Server (NTRS)
Lee, Byung Joon; Liou, Meng-Sing
2012-01-01
This paper describes the work for optimizing the propulsive efficiency of flapping airfoils, i.e., improving the thrust under constraining aerodynamic work during the flapping flights by changing their shape and trajectory of motion with the unsteady discrete adjoint approach. For unsteady problems, it is essential to properly resolving time scales of motion under consideration and it must be compatible with the objective sought after. We include both the instantaneous and time-averaged (periodic) formulations in this study. For the design optimization with shape parameters or motion parameters, the time-averaged objective function is found to be more useful, while the instantaneous one is more suitable for flow control. The instantaneous objective function is operationally straightforward. On the other hand, the time-averaged objective function requires additional steps in the adjoint approach; the unsteady discrete adjoint equations for a periodic flow must be reformulated and the corresponding system of equations solved iteratively. We compare the design results from shape and trajectory optimizations and investigate the physical relevance of design variables to the flapping motion at on- and off-design conditions.
Spectral monodromy of non-self-adjoint operators
Phan, Quang Sang
2014-01-15
In the present paper, we build a combinatorial invariant, called the “spectral monodromy” from the spectrum of a single (non-self-adjoint) h-pseudodifferential operator with two degrees of freedom in the semi-classical limit. Our inspiration comes from the quantum monodromy defined for the joint spectrum of an integrable system of n commuting self-adjoint h-pseudodifferential operators, given by S. Vu Ngoc [“Quantum monodromy in integrable systems,” Commun. Math. Phys. 203(2), 465–479 (1999)]. The first simple case that we treat in this work is a normal operator. In this case, the discrete spectrum can be identified with the joint spectrum of an integrable quantum system. The second more complex case we propose is a small perturbation of a self-adjoint operator with a classical integrability property. We show that the discrete spectrum (in a small band around the real axis) also has a combinatorial monodromy. The main difficulty in this case is that we do not know the description of the spectrum everywhere, but only in a Cantor type set. In addition, we also show that the corresponding monodromy can be identified with the classical monodromy, defined by J. Duistermaat [“On global action-angle coordinates,” Commun. Pure Appl. Math. 33(6), 687–706 (1980)].