Search NASA⌕ Search

SEARCH · Search NASA

Results for “stochastic approximation”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 235 records · Page 13

Estimation and Analysis of Nonlinear Stochastic Systems

The algebraic and geometric structures of certain classes of nonlinear stochastic systems were exploited in order to obtain useful stability and estimation results. The class of bilinear stochastic systems (or linear systems with multiplicative noise) was discussed. The stochastic stability of bilinear systems driven by colored noise was considered. Approximate methods for obtaining sufficient conditions for the stochastic stability of bilinear systems evolving on general Lie groups were discussed. Two classes of estimation problems involving bilinear systems were considered. It was proved that, for systems described by certain types of Volterra series expansions or by certain bilinear equations evolving on nilpotent or solvable Lie groups, the optimal conditional mean estimator consists of a finite dimensional nonlinear set of equations. The theory of harmonic analysis was used to derive suboptimal estimators for bilinear systems driven by white noise which evolve on compact Lie groups or homogeneous spaces.

Marcus, S. I.↗

Stochastic Programming Approach for Resource Selection under Demand Uncertainty

Cost-efficient selection and scheduling of a subset of geographically distributed resources to meet the demands of a scientific workflow is a challenging problem. The problem is exacerbated by uncertainties in demand and availability of resources. In this paper, we present a stochastic optimization based framework for robust decision making in the selection of distributed resources over a planning horizon under demand uncertainty. We present a novel two-stage stochastic programming model for resource selection, and implement an L-shaped decomposition algorithm to solve this model. A Sample Average Approximation algorithm is integrated to enable stochastic optimization to solve problems with a large number of scenarios. Using the metric of stochastic solution, we demonstrate up to 30% cost reduction relative to solutions without explicit consideration of demand uncertainty for a 24-month problem. We also demonstrate up to 54% cost reduction relative to a previously developed solution for a 36-month problem. We further argue that the composition of resources selected is superior to solutions computed without explicit consideration of uncertainties. Given the importance of resource selection and scheduling of complex scientific workflows, especially in the context of commercial cloud computing, we believe that our novel stochastic programming framework will benefit many researchers as well as users of distributed computing resources.

scientific workflows, task scheduling, optimizatio↗

Phase diagram of the hard-core Bose-Hubbard model on a checkerboard superlattice

We obtain the complete phase diagram of the hard-core Bose-Hubbard model in the presence of a period-two superlattice in two and three dimensions. First we acquire the phase boundaries between the superfluid phase and the 'trivial' insulating phases of the model (the completely-empty and completely-filled lattices) analytically. Next, the boundary between the superfluid phase and the half-filled Mott-insulating phase is obtained numerically, using the stochastic series expansion algorithm followed by finite-size scaling. We also compare our numerical results against the predictions of several approximation schemes, including two mean-field approaches and a fourth-order strong-coupling expansion, where we show that the latter method in particular is successful in producing an accurate picture of the phase diagram. Finally, we examine the extent to which several approximation schemes, such as the random phase approximation and the strong-coupling expansion, give an accurate description of the momentum distribution of the bosons inside the insulating phases.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Optimization Testbed Cometboards Extended into Stochastic Domain

COMparative Evaluation Testbed of Optimization and Analysis Routines for the Design of Structures (CometBoards) is a multidisciplinary design optimization software. It was originally developed for deterministic calculation. It has now been extended into the stochastic domain for structural design problems. For deterministic problems, CometBoards is introduced through its subproblem solution strategy as well as the approximation concept in optimization. In the stochastic domain, a design is formulated as a function of the risk or reliability. Optimum solution including the weight of a structure, is also obtained as a function of reliability. Weight versus reliability traced out an inverted-S-shaped graph. The center of the graph corresponded to 50 percent probability of success, or one failure in two samples. A heavy design with weight approaching infinity could be produced for a near-zero rate of failure that corresponded to unity for reliability. Weight can be reduced to a small value for the most failure-prone design with a compromised reliability approaching zero. The stochastic design optimization (SDO) capability for an industrial problem was obtained by combining three codes: MSC/Nastran code was the deterministic analysis tool, fast probabilistic integrator, or the FPI module of the NESSUS software, was the probabilistic calculator, and CometBoards became the optimizer. The SDO capability requires a finite element structural model, a material model, a load model, and a design model. The stochastic optimization concept is illustrated considering an academic example and a real-life airframe component made of metallic and composite materials.

Patnaik, Surya N.↗

A weighted ℓ{sub 1}-minimization approach for sparse polynomial chaos expansions

This work proposes a method for sparse polynomial chaos (PC) approximation of high-dimensional stochastic functions based on non-adapted random sampling. We modify the standard ℓ{sub 1}-minimization algorithm, originally proposed in the context of compressive sampling, using a priori information about the decay of the PC coefficients, when available, and refer to the resulting algorithm as weightedℓ{sub 1}-minimization. We provide conditions under which we may guarantee recovery using this weighted scheme. Numerical tests are used to compare the weighted and non-weighted methods for the recovery of solutions to two differential equations with high-dimensional random inputs: a boundary value problem with a random elliptic operator and a 2-D thermally driven cavity flow with random boundary condition.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

FuseIM: Fusing Probabilistic Traversals for Influence Maximization on Exascale Systems

Probabilistic breadth-first traversals (BPTs) are used in many network science and graph machine learning applications. In this paper, we are motivated by the application of BPTs in stochastic diffusion-based graph problems such as influence maximization. These applications heavily rely on BPTs to implement a Monte-Carlo sampling step for their approximations. Given the large sampling complexity, stochasticity of the diffusion process, and the inherent irregularity in real-world graph topologies, efficiently parallelizing these BPTs remains significantly challenging. In this paper, we present a new algorithm to fuse massive number of concurrently executing BPTs with random starts on the input graph. Our algorithm is designed to fuse BPTs by combining separate traversals into a unified frontier on distributed multi-GPU systems. To show the general applicability of the fused BPT technique, we have incorporated it into two state-of-the-art influence maximization parallel implementations (gIM and Ripples). Our experiments on up to 4K nodes of the OLCF Frontier supercomputer (32,768 GPUs and 196K CPU cores) show strong scaling behavior, and that fused BPTs can improve the performance of these implementations up to 34x (for gIM) and ~360x (for Ripples).

Neff, Reece W.↗

Stochastic Formulation of the Resolution of Identity: Application to Second Order Møller–Plesset Perturbation Theory

A stochastic orbital approach to the resolution of identity (RI) approximation for 4-index electron repulsion integrals (ERIs) is presented. The stochastic RI-ERIs are then applied to second order Møller–Plesset perturbation theory (MP2) utilizing a multiple stochastic orbital approach. The introduction of multiple stochastic orbitals results in an O(N AO 3 ) scaling for both the stochastic RI-ERIs and stochastic RI-MP2, NAO being the number of basis functions. For a range of water clusters we demonstrate that this method exhibits a small prefactor and observed scalings of O(N e 2.4 ) for total energies and O(N e 3.1 ) for forces (N e being the number of correlated electrons), outperforming MP2 for clusters with as few as 21 water molecules.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Nonlinear simulations to optimize magnetic nanoparticle hyperthermia

Magnetic nanoparticle hyperthermia is an attractive emerging cancer treatment, but the acting microscopic energy deposition mechanisms are not well understood and optimization suffers. We describe several approximate forms for the characteristic time of Néel rotations with varying properties and external influences. We then present stochastic simulations that show agreement between the approximate expressions and the micromagnetic model. The simulations show nonlinear imaginary responses and associated relaxational hysteresis due to the field and frequency dependencies of the magnetization. This suggests that efficient heating is possible by matching fields to particles instead of resorting to maximizing the power of the applied magnetic fields.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Interparameter trade-off quantification and reduction in isotropic-elastic full-waveform inversion: synthetic experiments and Hussar land data set application

The problem of inverting for multiple physical parameters in the subsurface using seismic full-waveform inversion (FWI) is complicated by interparameter trade-off arising from inherent ambiguities between different physical parameters. Parameter resolution is often characterized using scattering radiation patterns, but these neglect some important aspects of interparameter trade-off. More general analysis and mitigation of interparameter trade-off in isotropic-elastic FWI is possible through judiciously chosen multiparameter Hessian matrix-vector products. We show that products of multiparameter Hessian off-diagonal blocks with model perturbation vectors, referred to as interparameter contamination kernels , are central to the approach. We then apply the multiparameter Hessian to various vectors designed to provide information regarding the strengths and characteristics of interparameter contamination, both locally and within the whole volume. With numerical experiments, we observe that S-wave velocity perturbations introduce strong contaminations into density and phase-reversed contaminations into P-wave velocity, but themselves experience only limited contaminations from other parameters. Based on these findings, we introduce a novel strategy to mitigate the influence of interparameter trade-off with approximate contamination kernels. Furthermore, we recommend that the local spatial and interparameter trade-off of the inverted models be quantified using extended multiparameter point spread functions (EMPSFs) obtained with pre-conditioned conjugate-gradient algorithm. Compared to traditional point spread functions, the EMPSFs appear to provide more accurate measurements for resolution analysis, by de-blurring the estimations, scaling magnitudes and mitigating interparameter contamination. Approximate eigenvalue volumes constructed with stochastic probing approach are proposed to evaluate the resolution of the inverted models within the whole model. With a synthetic Marmousi model example and a land seismic field data set from Hussar, Alberta, Canada, we confirm that the new inversion strategy suppresses the interparameter contamination effectively and provides more reliable density estimations in isotropic-elastic FWI as compared to standard simultaneous inversion approach.

58 GEOSCIENCES↗

Analysis of bilinear stochastic systems

Analysis of stochastic dynamical systems that involve multiplicative (bilinear) noise processes. After defining the systems of interest, consideration is given to the evolution of the moments of such systems, the question of stochastic stability, and estimation for bilinear stochastic systems. Both exact and approximate methods of analysis are introduced, and, in particular, the uses of Lie-theoretic concepts and harmonic analysis are discussed.

Willsky, A. S.↗

Stochastic estimation of organized structures in turbulent channel flow

Linear stochastic estimation is used to approximate the conditional vector fields associated with high Reynolds stress producing events in numerically simulated turbulent channel flow. Joint probability density distributions of u-v are presented for y(+) between 1.35 and 180, and at each position the u-v values that make the greatest contributions to the average uv3 are used to define conditional ejection- and sweep-type events. In the y-z plane the conditional fields appear to be pairs of counter-rotating vortices. When the conditional event defined by the weighted probability density function is specified at a point close to the wall, the spacing and size of these eddies are consistent with the accepted wall streak spacing. As this point moves outwards, the eddy size increases significantly. An abrupt change in the flow angle occurs in the buffer layer and may indicate transition from streamwise oriented wall layer structures to hairpin vortices characterizing the outer layer.

Moin, P.↗

Modern CACSD using the Robust-Control Toolbox

The Robust-Control Toolbox is a collection of 40 M-files which extend the capability of PC/PRO-MATLAB to do modern multivariable robust control system design. Included are robust analysis tools like singular values and structured singular values, robust synthesis tools like continuous/discrete H(exp 2)/H infinity synthesis and Linear Quadratic Gaussian Loop Transfer Recovery methods and a variety of robust model reduction tools such as Hankel approximation, balanced truncation and balanced stochastic truncation, etc. The capabilities of the toolbox are described and illustated with examples to show how easily they can be used in practice. Examples include structured singular value analysis, H infinity loop-shaping and large space structure model reduction.

Chiang, Richard Y.↗

An X-Ray Study of Grain Evolution behind Non-Radiative Shock

We model the dynamics, the destruction by sputtering, and the infrared (IR) emission from collisionally heated dust grains in fast (>/= 400 km s(exp -1)) astrophysical shocks in order to develop IR diagnostics for the destruction of grains in these environments. The calculations take into account the feedback from sputtering and IR emission on the gas-phase abundances, the cooling, and the ionization and thermal structure of the shock. Sputtering changes the initial grain size distribution, creating a deficiency of small (radius <50 A) grains compared to their preshock abundances. The altered grain size distribution depends on shock velocity and the density of the interstellar medium. Dust particles with sizes below approximately equal 300 A are stochastically heated, undergo temperature fluctuations, and radiate an excess of near-infrared emission (i </= 40 (micron)meter) over that expected for grains in thermal equilibrium. This near-infrared excess is a measure of the abundance of small grains and therefore a powerful diagnostic for the amount of destruction the grains were subjected to in the shock. We present here IR spectra from collisionally heated dust for a variety of shocks, and depict the changes in the spectra as a function of postshock column density. Our studies compliment those of Vancura et at. that examined the effects of the release of the sputtered refractory elements on the ultraviolet and X-ray emission. Multiwavelength observations at X-ray, Ultraviolet (UV), and IR wavelengths are therefore essential in piecing together a comprehensive picture of the physics of grain destruction in fast astrophysical shocks.

Vancura, Olaf↗

Initialization and Restart in Stochastic Local Search: Computing a Most Probable Explanation in Bayesian Networks

For hard computational problems, stochastic local search has proven to be a competitive approach to finding optimal or approximately optimal problem solutions. Two key research questions for stochastic local search algorithms are: Which algorithms are effective for initialization? When should the search process be restarted? In the present work we investigate these research questions in the context of approximate computation of most probable explanations (MPEs) in Bayesian networks (BNs). We introduce a novel approach, based on the Viterbi algorithm, to explanation initialization in BNs. While the Viterbi algorithm works on sequences and trees, our approach works on BNs with arbitrary topologies. We also give a novel formalization of stochastic local search, with focus on initialization and restart, using probability theory and mixture models. Experimentally, we apply our methods to the problem of MPE computation, using a stochastic local search algorithm known as Stochastic Greedy Search. By carefully optimizing both initialization and restart, we reduce the MPE search time for application BNs by several orders of magnitude compared to using uniform at random initialization without restart. On several BNs from applications, the performance of Stochastic Greedy Search is competitive with clique tree clustering, a state-of-the-art exact algorithm used for MPE computation in BNs.

Mengshoel, Ole J.↗

Analysis of stochastically forced quasi-periodic attractors

A problem of the analysis of stochastically forced quasi-periodic auto-oscillations of nonlinear dynamic systems is considered. A stationary distribution of random trajectories in the neighborhood of the corresponding deterministic attractor (torus) is studied. A parametric description of quadratic approximation of the quasipotential based on the stochastic sensitivity functions (SSF) technique is given. Using this technique, we analyse a dispersion of stochastic flows near the torus. For the case of two-torus in three-dimensional space, the stochastic sensitivity function is constructed.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Improved statistical models for limited datasets in uncertainty quantification using stochastic collocation

This paper presents a data-driven framework for performing uncertainty quantification (UQ) by choosing a stochastic model that accurately describes the sources of uncertainty in a system. This model is propagated through an appropriate response surface function that approximates the behavior of this system using stochastic collocation. Given a sample of data describing the uncertainty in the inputs, our goal is to estimate a probability density function (PDF) using the kernel moment matching (KMM) method so that this PDF can be used to accurately reproduce statistics like mean and variance of the response surface function. Instead of constraining the PDF to be optimal for a particular response function, we show that we can use the properties of stochastic collocation to make the estimated PDF optimal for a wide variety of response functions. We contrast this method with other traditional procedures that rely on the Maximum Likelihood approach, like kernel density estimation (KDE) and its adaptive modification (AKDE). We argue that this modified KMM method tries to preserve what is known from the given data and is the better approach when the available data is limited in quantity. We test the performance of these methods for both univariate and multivariate density estimation by sampling random datasets from known PDFs and then measuring the accuracy of the estimated PDFs, using the known PDF as a reference. Comparing the output mean and variance estimated with the empirical moments using the raw data sample as well as the actual moments using the known PDF, we show that the KMM method performs better than KDE and AKDE in predicting these moments with greater accuracy. This improvement in accuracy is also demonstrated for the case of UQ in electrostatic and electrothermomechanical microactuators. We show how our framework results in the accurate computation of statistics in micromechanical systems.

97 MATHEMATICS AND COMPUTING↗

Goal-oriented sensitivity analysis for lattice kinetic Monte Carlo simulations

In this paper we propose a new class of coupling methods for the sensitivity analysis of high dimensional stochastic systems and in particular for lattice Kinetic Monte Carlo (KMC). Sensitivity analysis for stochastic systems is typically based on approximating continuous derivatives with respect to model parameters by the mean value of samples from a finite difference scheme. Instead of using independent samples the proposed algorithm reduces the variance of the estimator by developing a strongly correlated-“coupled”- stochastic process for both the perturbed and unperturbed stochastic processes, defined in a common state space. The novelty of our construction is that the new coupled process depends on the targeted observables, e.g., coverage, Hamiltonian, spatial correlations, surface roughness, etc., hence we refer to the proposed method as goal-oriented sensitivity analysis. In particular, the rates of the coupled Continuous Time Markov Chain are obtained as solutions to a goal-oriented optimization problem, depending on the observable of interest, by considering the minimization functional of the corresponding variance. We show that this functional can be used as a diagnostic tool for the design and evaluation of different classes of couplings. Furthermore, the resulting KMC sensitivity algorithm has an easy implementation that is based on the Bortz–Kalos–Lebowitz algorithm's philosophy, where events are divided in classes depending on level sets of the observable of interest. Finally, we demonstrate in several examples including adsorption, desorption, and diffusion Kinetic Monte Carlo that for the same confidence interval and observable, the proposed goal-oriented algorithm can be two orders of magnitude faster than existing coupling algorithms for spatial KMC such as the Common Random Number approach. We also provide a complete implementation of the proposed sensitivity analysis algorithms, including various spatial KMC examples, in a supplementary MATLAB source code.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗