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 181 records · Page 10

Heavy-ion collisions and fission dynamics with the time-dependent Hartree–Fock theory and its extensions

Microscopic methods and tools to describe nuclear dynamics have considerably been improved in the past few years. They are based on the time-dependent Hartree–Fock (TDHF) theory and its extensions to include pairing correlations and quantum fluctuations. The TDHF theory is the lowest level of approximation of a range of methods to solve the quantum many-body problem, showing its universality to describe many-fermion dynamics at the mean-field level. The range of applications of TDHF to describe realistic systems allowing for detailed comparisons with experiment has considerably increased. For instance, TDHF is now commonly used to investigate fusion, multi-nucleon transfer and quasi-fission reactions. Thanks to the inclusion of pairing correlations, it has also recently led to breakthroughs in our description of the saddle to scission evolution, and, in particular, the non-adiabatic effects near scission. Beyond mean-field approaches such as the time-dependent random-phase approximation (TDRPA) and stochastic mean-field methods have reached the point where they can be used for realistic applications. We review recent progresses in both techniques and applications to heavy-ion collision and fission.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Low Mach number fluctuating hydrodynamics for electrolytes

Here, we formulate and study computationally the low Mach number fluctuating hydrodynamic equations for electrolyte solutions. We are also interested in studying transport in mixtures of charged species at the mesoscale, down to scales below the Debye length, where thermal fluctuations have a significant impact on the dynamics. Continuing our previous work on fluctuating hydrodynamics of multicomponent mixtures of incompressible isothermal miscible liquids (A. Donev, et al., Physics of Fluids, 27, 3, 2015), we now include the effect of charged species using a quasielectrostatic approximation. Localized charges create an electric field, which in turn provides additional forcing in the mass and momentum equations. Our low Mach number formulation eliminates sound waves from the fully compressible formulation and leads to a more computationally efficient quasi-incompressible formulation. Furthermore, we demonstrate our ability to model saltwater (NaCl) solutions in both equilibrium and nonequilibrium settings. We show that our algorithm is second-order in the deterministic setting, and for length scales much greater than the Debye length gives results consistent with an electroneutral/ambipolar approximation. In the stochastic setting, our model captures the predicted dynamics of equilibrium and nonequilibrium fluctuations. We also identify and model an instability that appears when diffusive mixing occurs in the presence of an applied electric field.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

WE-H-207A-09: Theoretical Limits to Molecular Biomarker Detection Using Magnetic Nanoparticles

Purpose: Estimate the limits of molecular biomarker detection using magnetic nanoparticle methods like in vivo ELISA. Methods: Magnetic nanoparticles in an alternating magnetic field produce a magnetization that can be detected at exceedingly low levels because the signal at the harmonic frequencies is uniquely produced by the nanoparticles. Because the magnetization can also be used to characterize the nanoparticle rotational freedom, the bound state can be found. If the nanoparticles are coated with molecules that bind the desired biomarker, the rotational freedom reflects the biomarker concentration. The irreducible noise limit is the thermal noise or Johnson noise of the tissue and the contrast that can be measured must be larger than that limit. The contrast produced is a function of the applied field and depends strongly on nanoparticle volume. We have estimated the contrast using a Langevin function of a single composite variable to approximate the full stochastic Langevin equation for nanoparticle dynamics. Results: The thermal noise for a bandwidth reasonable for spectroscopy suggests mid zeptomolar (10–21) to low attomolar (10–18) concentrations can be measured in a volume that is 10cm in scale. The suggested sensitivity is far below the physiologically concentrations of almost all critical biomarkers including cytokines (picomolar), hormones (nanomolar) and heat shock proteins. Conclusion: The sensitivity of in vivo ELISA concentration measurements should be sufficient to measure physiological concentrations of critical biomarkers like cytokines in vivo. Further the sensitivity should be sufficient to measure concentrations of other biomarkers that are six to eight orders of magnitude lower in concentration than immune signaling molecules like cytokines. NIH - 1U54CA151662-01 Department of Radiology.

60 APPLIED LIFE SCIENCES↗

Tradeoff studies in multiobjective insensitive design of airplane control systems

A computer aided design method for multiobjective parameter-insensitive design of airplane control systems is described. Methods are presented for trading off nominal values of design objectives against sensitivities of the design objectives to parameter uncertainties, together with guidelines for designer utilization of the methods. The methods are illustrated by application to the design of a lateral stability augmentation system for two supersonic flight conditions of the Shuttle Orbiter. Objective functions are conventional handling quality measures and peak magnitudes of control deflections and rates. The uncertain parameters are assumed Gaussian, and numerical approximations of the stochastic behavior of the objectives are described. Results of applying the tradeoff methods to this example show that stochastic-insensitive designs are distinctly different from deterministic multiobjective designs. The main penalty for achieving significant decrease in sensitivity is decreased speed of response for the nominal system.

Schy, A. A.↗

Dual adaptive control: Design principles and applications

The design of an actively adaptive dual controller based on an approximation of the stochastic dynamic programming equation for a multi-step horizon is presented. A dual controller that can enhance identification of the system while controlling it at the same time is derived for multi-dimensional problems. This dual controller uses sensitivity functions of the expected future cost with respect to the parameter uncertainties. A passively adaptive cautious controller and the actively adaptive dual controller are examined. In many instances, the cautious controller is seen to turn off while the latter avoids the turn-off of the control and the slow convergence of the parameter estimates, characteristic of the cautious controller. The algorithms have been applied to a multi-variable static model which represents a simplified linear version of the relationship between the vibration output and the higher harmonic control input for a helicopter. Monte Carlo comparisons based on parametric and nonparametric statistical analysis indicate the superiority of the dual controller over the baseline controller.

Mookerjee, Purusottam↗

Multi-level methods and approximating distribution functions

Biochemical reaction networks are often modelled using discrete-state, continuous-time Markov chains. System statistics of these Markov chains usually cannot be calculated analytically and therefore estimates must be generated via simulation techniques. There is a well documented class of simulation techniques known as exact stochastic simulation algorithms, an example of which is Gillespie’s direct method. These algorithms often come with high computational costs, therefore approximate stochastic simulation algorithms such as the tau-leap method are used. However, in order to minimise the bias in the estimates generated using them, a relatively small value of tau is needed, rendering the computational costs comparable to Gillespie’s direct method. The multi-level Monte Carlo method (Anderson and Higham, Multiscale Model. Simul. 10:146–179, 2012) provides a reduction in computational costs whilst minimising or even eliminating the bias in the estimates of system statistics. This is achieved by first crudely approximating required statistics with many sample paths of low accuracy. Then correction terms are added until a required level of accuracy is reached. Recent literature has primarily focussed on implementing the multi-level method efficiently to estimate a single system statistic. However, it is clearly also of interest to be able to approximate entire probability distributions of species counts. We present two novel methods that combine known techniques for distribution reconstruction with the multi-level method. We demonstrate the potential of our methods using a number of examples.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Multi-fidelity stochastic collocation method for computation of statistical moments

We present an efficient numerical algorithm to approximate the statistical moments of stochastic problems, in the presence of models with different fidelities. The method extends the multi-fidelity approximation method developed in . By combining the efficiency of low-fidelity models and the accuracy of high-fidelity models, our method exhibits fast convergence with a limited number of high-fidelity simulations. We establish an error bound of the method and present several numerical examples to demonstrate the efficiency and applicability of the multi-fidelity algorithm.

97 MATHEMATICS AND COMPUTING↗

A Low-Rank Solver for the Navier--Stokes Equations with Uncertain Viscosity

In this work, we study an iterative low-rank approximation method for the solution of the steady-state stochastic Navier--Stokes equations with uncertain viscosity. The method is based on linearization schemes using Picard and Newton iterations and stochastic finite element discretizations of the linearized problems. For computing the low-rank approximate solution, we adapt the nonlinear iterations to an inexact and low-rank variant, where the solution of the linear system at each nonlinear step is approximated by a quantity of low rank. This is achieved by using a tensor variant of the GMRES method as a solver for the linear systems. We explore the inexact low-rank nonlinear iteration with a set of benchmark problems, using a model of flow over an obstacle, under various configurations characterizing the statistical features of the uncertain viscosity, and we demonstrate its effectiveness by extensive numerical experiments.

97 MATHEMATICS AND COMPUTING↗

Lignin valorization process control under feedstock uncertainty through a dynamic stochastic programming approach

The randomness introduced by reactants is an issue when processing renewable bioresources. In this paper, we apply tools from dynamic stochastic programming theory to the biochemical process. Instead of introducing extra preprocessing units, we consider the inherent randomness of the process and optimize in expectation the performance of the system. In a general setting, this is a multistage stochastic optimization problem and we investigate its approximate solution via two approaches, namely stochastic dual dynamic programming (SDDP) and the finite state, finite action Markov decision process (MDP) framework. These two methods are implemented to a case study of lignin valorization that is crucial to a cost-effective biorefinery process using biomass as the feedstock.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Worldline casting of the stochastic vacuum model and nonperturbative properties of QCD: General formalism and applications

The stochastic vacuum model for QCD, proposed by Dosch and Simonov, is fused with a worldline casting of the underlying theory, i.e. QCD. Important nonperturbative features of the model are studied. In particular, contributions associated with the spin-field interaction are calculated, and the validity of both the loop equations and of the Bianchi identity is explicitly demonstrated. As an application, a simulated meson-meson scattering problem is studied in the Regge kinematical regime. The process is modeled in terms of the helicoidal Wilson contour along the lines introduced by Janik and Peschanski in a related study based on an AdS/CFT-type approach. Working strictly in the framework of the stochastic vacuum model and in a semiclassical approximation scheme, the Regge behavior for the scattering amplitude is demonstrated. Going beyond this approximation, the contribution resulting from boundary fluctuations of the Wilson loop contour is also estimated.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Solution of the stochastic control problem in unbounded domains.

Bellman's dynamic programming equation for the optimal index and control law for stochastic control problems is a parabolic or elliptic partial differential equation frequently defined in an unbounded domain. Existing methods of solution require bounded domain approximations, the application of singular perturbation techniques or Monte Carlo simulation procedures. In this paper, using the fact that Poisson impulse noise tends to a Gaussian process under certain limiting conditions, a method which achieves an arbitrarily good approximate solution to the stochastic control problem is given. The method uses the two iterative techniques of successive approximation and quasi-linearization and is inherently more efficient than existing methods of solution.

Robinson, P.↗

Switching of bound vector solitons for the coupled nonlinear Schroedinger equations with nonhomogenously stochastic perturbations

We investigate the dynamics of the bound vector solitons (BVSs) for the coupled nonlinear Schroedinger equations with the nonhomogenously stochastic perturbations added on their dispersion terms. Soliton switching (besides soliton breakup) can be observed between the two components of the BVSs. Rate of the maximum switched energy (absolute values) within the fixed propagation distance (about 10 periods of the BVSs) enhances in the sense of statistics when the amplitudes of stochastic perturbations increase. Additionally, it is revealed that the BVSs with enhanced coherence are more robust against the perturbations with nonhomogenous stochasticity. Diagram describing the approximate borders of the splitting and non-splitting areas is also given. Our results might be helpful in dynamics of the BVSs with stochastic noises in nonlinear optical fibers or with stochastic quantum fluctuations in Bose-Einstein condensates.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Exact and inexact subsampled Newton methods for optimization

Abstract The paper studies the solution of stochastic optimization problems in which approximations to the gradient and Hessian are obtained through subsampling. We first consider Newton-like methods that employ these approximations and discuss how to coordinate the accuracy in the gradient and Hessian to yield a superlinear rate of convergence in expectation. The second part of the paper analyzes an inexact Newton method that solves linear systems approximately using the conjugate gradient (CG) method, and that samples the Hessian and not the gradient (the gradient is assumed to be exact). We provide a complexity analysis for this method based on the properties of the CG iteration and the quality of the Hessian approximation, and compare it with a method that employs a stochastic gradient iteration instead of the CG method. We report preliminary numerical results that illustrate the performance of inexact subsampled Newton methods on machine learning applications based on logistic regression.

Mathematics↗

Stochastic GW -GPU: Rapid Quasi-Particle Energies for Molecules beyond 10,000 Atoms

StochasticGW is a code for computing accurate quasi-particle (QP) energies of molecules and material systems in the GW approximation. StochasticGW utilizes the stochastic Resolution of the Identity (sROI) technique to enable a massively parallel implementation with computational costs that scale semilinearly with system size, allowing the method to access systems with tens of thousands of electrons. Here, we introduce a new implementation, StochasticGW-GPU, for which the main bottleneck steps have been ported to GPUs and give substantial performance improvements over previous versions of the code. We showcase the new code by computing band gaps of hydrogenated silicon clusters (Si x H y ) containing up to 10,001 atoms and 35,144 electrons, and we obtain individual QP energies with a statistical precision of better than ±0.03 eV with times-to-solution of less than 1 h.

Thomas, Phillip S. [Lawrence Berkeley National Lab↗

A parton picture of de Sitter space during slow-roll inflation

It is well-known that expectation values in de Sitter space are afflicted by infra-red divergences. Long ago, Starobinsky proposed that infra-red effects in de Sitter space could be accommodated by evolving the long-wavelength part of the field according to the classical equations of motion plus a stochastic source term. I argue that - when quantum-mechanical loop corrections are taken into account - the separate universe picture of superhorizon evolution in de Sitter space is equivalent, in a certain leading-logarithm approximation, to Starobinsky's stochastic approach. In particular the time evolution of a box of de Sitter space can be understood in exact analogy with the DGLAP evolution of partons within a hadron, which describes a slow logarithmic evolution in the distribution of the hadron's constituent partons with the energy scale at which they are probed.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

An efficient computational method for solving nonlinear stochastic Itô integral equations: Application for stochastic problems in physics

Because of the nonlinearity, closed-form solutions of many important stochastic functional equations are virtually impossible to obtain. Thus, numerical solutions are a viable alternative. In this paper, a new computational method based on the generalized hat basis functions together with their stochastic operational matrix of Itô-integration is proposed for solving nonlinear stochastic Itô integral equations in large intervals. In the proposed method, a new technique for computing nonlinear terms in such problems is presented. The main advantage of the proposed method is that it transforms problems under consideration into nonlinear systems of algebraic equations which can be simply solved. Error analysis of the proposed method is investigated and also the efficiency of this method is shown on some concrete examples. The obtained results reveal that the proposed method is very accurate and efficient. As two useful applications, the proposed method is applied to obtain approximate solutions of the stochastic population growth models and stochastic pendulum problem.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗