Search NASA⌕ Search

SEARCH · Search NASA

Results for “Stochastic optimization”

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 1,351 records · Page 75

Asymmetric Dual Sensor Latency in Distributed Control

Considerable geographical separation between various phasor measurement unit sensor suites used in the wide-area damping control architecture results in time delays with varying steady-state and stochastic components. The consequences of asymmetric latent sensors used in a standard feedback configuration are presented, indicating that asymmetry of the delay provides not only a previously suspected modification to the argument of the loop transmission function, but also a potentially substantial impact on its modulus. Under particular delay configurations, the controller may experience the introduction of varying crossovers in the loop transmission, resulting in a shifting of control bandwidth and previously unconsidered (in)stability characteristics (e.g., alteration of relative stability margins). With consideration to the theoretical results provided, the optimal sensor location based on delay characteristics is provided. Finally, results of the study indicate that it is most frequently best practice to act on the most current sample in the feedback channel, rather than wait for precise time alignment of the delayed sensor.

42 ENGINEERING↗

Report of the AAPM Task Group No. 105: Issues associated with clinical implementation of Monte Carlo-based photon and electron external beam treatment planning

The Monte Carlo (MC) method has been shown through many research studies to calculate accurate dose distributions for clinical radiotherapy, particularly in heterogeneous patient tissues where the effects of electron transport cannot be accurately handled with conventional, deterministic dose algorithms. Despite its proven accuracy and the potential for improved dose distributions to influence treatment outcomes, the long calculation times previously associated with MC simulation rendered this method impractical for routine clinical treatment planning. However, the development of faster codes optimized for radiotherapy calculations and improvements in computer processor technology have substantially reduced calculation times to, in some instances, within minutes on a single processor. These advances have motivated several major treatment planning system vendors to embark upon the path of MC techniques. Several commercial vendors have already released or are currently in the process of releasing MC algorithms for photon and/or electron beam treatment planning. Consequently, the accessibility and use of MC treatment planning algorithms may well become widespread in the radiotherapy community. With MC simulation, dose is computed stochastically using first principles; this method is therefore quite different from conventional dose algorithms. Issues such as statistical uncertainties, the use of variance reduction techniques, the ability to account for geometric details in the accelerator treatment head simulation, and other features, are all unique components of a MC treatment planning algorithm. Successful implementation by the clinical physicist of such a system will require an understanding of the basic principles of MC techniques. The purpose of this report, while providing education and review on the use of MC simulation in radiotherapy planning, is to set out, for both users and developers, the salient issues associated with clinical implementation and experimental verification of MC dose algorithms. As the MC method is an emerging technology, this report is not meant to be prescriptive. Rather, it is intended as a preliminary report to review the tenets of the MC method and to provide the framework upon which to build a comprehensive program for commissioning and routine quality assurance of MC-based treatment planning systems.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Parameter estimation by implicit sampling

© 2015 Mathematical Sciences Publishers. Implicit sampling is a weighted sampling method that is used in data assimilation to sequentially update state estimates of a stochastic model based on noisy and incomplete data. Here we apply implicit sampling to sample the posterior probability density of parameter estimation problems. The posterior probability combines prior information about the parameter with information from a numerical model, e.g., a partial differential equation (PDE), and noisy data. The result of our computations are parameters that lead to simulations that are compatible with the data. We demonstrate the usefulness of our implicit sampling algorithm with an example from subsurface flow. For an efficient implementation, we make use of multiple grids, BFGS optimization coupled to adjoint equations, and Karhunen-Loève expansions for dimensional reduction. Several difficulties of Markov chain Monte Carlo methods, e.g., estimation of burn-in times or correlations among the samples, are avoided because the implicit samples are independent.

Morzfeld, Matthias↗

Economic Dispatch for Microgrid Containing Electric Vehicles via Probabilistic Modeling: Preprint

In this paper, an economic dispatch model with probabilistic modeling is developed for a microgrid. The electric power supply in a microgrid consists of conventional power plants and renewable energy power plants, such as wind and solar power plants. Because of the fluctuation in the output of solar and wind power plants, an empirical probabilistic model is developed to predict their hourly output. According to different characteristics of wind and solar power plants, the parameters for probabilistic distribution are further adjusted individually for both. On the other hand, with the growing trend in plug-in electric vehicles (PHEVs), an integrated microgrid system must also consider the impact of PHEVs. The charging loads from PHEVs as well as the discharging output via the vehicle-to-grid (V2G) method can greatly affect the economic dispatch for all of the micro energy sources in a microgrid. This paper presents an optimization method for economic dispatch in a microgrid considering conventional power plants, renewable power plants, and PHEVs. The simulation results reveal that PHEVs with V2G capability can be an indispensable supplement in a modern microgrid.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Addressing Load Imbalance in Bioinformatics and Biomedical Applications: Efficient Scheduling across Multiple GPUs

Computational bioinformatics and biomedical applications frequently contain heterogeneously sized units of work or tasks, for instance due to variability in the sizes of biological sequences and molecules. Variable-sized workloads lead to load imbalances in parallel implementations which detract from efficiency and performance. Many modern computing resources now have multiple graphics processing units(GPUs) per computer for acceleration. These multiple GPU resources need to be used efficiently through balancing of workloads across the GPUs. OpenMP is a portable directive-based parallel programming API used ubiquitously in bioscience applications to program CPUs; recently, the use of OpenMP directives for GPU acceleration has become possible. Here, motivated by experiences with imbalanced loads in GPU-accelerated bioinformatics applications, we address the load balancing problem using OpenMP task-to-GPU scheduling combined with OpenMP GPU offloading for multiply heterogeneous workloads – loads with both variable input sizes, and simultaneously, variable convergence rates for algorithms with a stochastic component – scheduled across multiple GPUs. We aim to develop strategies which are both easy to use and have lower overheads, and may be incorporated incrementally in existing programs which already make use of OpenMP for CPU-based threading in order to make use of multi-GPU computers. We test different combinations of input size variability and convergence rate variability, and characterize the effects of these different scenarios on the performance of scheduling strategies across multiple GPUs with OpenMP. We present several dynamic scheduling solutions for different parallel patterns, explore optimizations, and provide publicly available example computational kernels to make these strategies easy to use in programs. This work will enable application developers to efficiently and easily use multiple GPUs for imbalanced workloads found in bioinformatics and biomedical applications.

Thavappiragasam, Mathialakan↗

Finding the Lowest-Energy Crystal Structure Starting From Randomly Selected Lattice Vectors and Atomic Positions: First Principles Evolutionary Study of the Au-Pd, Cd-Pt, Al-Sc, Cu-Pd, Pd-Ti, and Ir-N Binary Systems

Two types of global space-group optimization (GSGO) problems can be recognized in binary metallic alloys A{sub q}B{sub 1-q}: (1) configuration search problems, where the underlying crystal lattice is known and the aim is finding the most favorable decoration of the lattice by A and B atoms and (2) lattice-type search problems, where neither the lattice type nor the decorations are given and the aim is finding energetically favorable lattice vectors and atomic occupations. Here, we address the second, lattice-type search problem in binary A{sub q}B{sub 1-q} metallic alloys, where the constituent solids A and B have different lattice types. We tackle this GSGO problem using an evolutionary algorithm, where a set of crystal structures with randomly selected lattice vectors and site occupations is evolved through a sequence of generations in which a given number of structures of highest LDA energy are replaced by new ones obtained by the generational operations of mutation or mating. Each new structure is locally relaxed to the nearest total-energy minimum by using the ab initio atomic forces and stresses. We applied this first-principles evolutionary GSGO scheme to metallic alloy systems where the nature of the intermediate A-B compounds is difficult to guess either because pure A and pure B have different lattice types and the (1) intermediate compound has the structure of one end-point (Al{sub 3}Sc, AlSc{sub 3}, CdPt{sub 3}), or (2) none of them (CuPd, AlSc), or (3) when the intermediate compound has lattice sites belonging simultaneously to a few types (fcc, bcc) (PdTi{sub 3}). The method found the correct structures, L1{sub 2} type for Al{sub 3}Sc, D0{sub 19} type for AlSc3, 'CdPt{sub 3}' type for CdPt{sub 3}, B2 type for CuPd and AlSc, and A15 type for PdTi{sub 3}. However, in such stochastic methods, success is not guaranteed, since many independently started evolutionary sequences produce at the end different final structures: one has to select the lowest-energy result from a set of such independently started sequences. Interestingly, we also predict a hitherto unknown (P 2/m) structure of the hard compound IrN{sub 2} with energy lower than all previous predictions.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

A multitower measurement network estimate of California's methane emissions

In this paper, we present an analysis of methane (CH 4 ) emissions using atmospheric observations from five sites in California's Central Valley across different seasons (September 2010 to June 2011). CH 4 emissions for spatial regions and source sectors are estimated by comparing measured CH 4 mixing ratios with transport model (Weather Research and Forecasting and Stochastic Time-Inverted Lagrangian Transport) predictions based on two 0.1° CH 4 (seasonally varying “California-specific” (California Greenhouse Gas Emission Measurements, CALGEM) and a static global (Emission Database for Global Atmospheric Research, release version 42, EDGAR42)) prior emission models. Region-specific Bayesian analyses indicate that for California's Central Valley, the CALGEM- and EDGAR42-based inversions provide consistent annual total CH 4 emissions (32.87 ± 2.09 versus 31.60 ± 2.17 Tg CO 2 eq yr -1 ; 68% confidence interval (CI), assuming uncorrelated errors between regions). Summing across all regions of California, optimized CH 4 emissions are only marginally consistent between CALGEM- and EDGAR42-based inversions (48.35 ± 6.47 versus 64.97 ± 11.85 Tg CO 2 eq), because emissions from coastal urban regions (where landfill and natural gas emissions are much higher in EDGAR than CALGEM) are not strongly constrained by the measurements. Combining our results with those from a recent study of the South Coast Air Basin narrows the range of estimates to 43–57 Tg CO2eq yr -1 (1.3–1.8 times higher than the current state inventory). Finally, these results suggest that the combination of rural and urban measurements will be necessary to verify future changes in California's total CH 4 emissions.

54 ENVIRONMENTAL SCIENCES↗

Endogenizing Probabilistic Resource Adequacy Risks in Deterministic Capacity Expansion Models

In this work, we demonstrate how power system capacity expansion models can understate the stochastic effects of thermal outages when considering resource availabilities on an hourly expected value basis, yielding system designs with multiple orders of magnitude more shortfall risk than stated adequacy targets. We develop a novel approximation approach to efficiently endogenize awareness of this risk in a deterministic, linear capacity expansion framework. We compare this approach to exogenous tuning of an energy reserve margin, the leading alternative method to compensate for unmodeled probabilistic shortfall risk. Empirical results from a test system show that the new endogenous method cost-effectively meets all regional reliability targets with a single optimization solve, and produces a near-identical system design as the incumbent method without the need for repeated re-optimizations to find an appropriate reserve level. The endogenous method may also use iterative re-optimizations to further improve solution quality, although these incremental benefits were modest in the system studied.

capacity expansion modeling↗

Optimal Control of Differentially Private EV Charging: A Scalable Learning Approach Under Uncertainty

Internet of Things (IoT)-enabled electric vehicles (IoEVs) enable intelligent charging coordination that accounts for grid congestion. However, increased data exchange raises privacy concerns, as charging patterns can reveal sensitive driver behavior to grid operators. Here, we propose a differentially private (DP) EV charging framework that enables coordinated control while protecting driver data with theoretical privacy guarantees. Nevertheless, integrating DP inevitably introduces uncertainty into the control strategy for EVs, which can lead to infeasible solutions. To tackle this challenge, we develop a feasible and scalable control algorithm based on constrained reinforcement learning (CRL) and convex hulls. While our framework is designed to handle the uncertainty introduced by DP, it is general and also applicable to other sources of uncertainty in EV charging, such as the stochastic nature of driver behavior and renewable variability. This ensures feasible and privacy-preserving coordination of EV charging at scale. Our method constructs convex hulls within the action space to guarantee feasibility under stochastic constraints and incorporates constraint reduction techniques to improve scalability. Case studies based on IEEE benchmark systems demonstrate that the proposed approach effectively balances feasibility under uncertainty, scalability, and privacy in large-scale EV charging control.

Engineering - Power transmission and distribution↗

TH-CD-207B-06: Swank Factor of Segmented Scintillators in Multi-Slice CT Detectors: Pulse Height Spectra and Light Escape

Purpose: Pulse height spectra (PHS) have been used to determine the Swank factor of a scintillator by measuring fluctuations in its light output per x-ray interaction. The Swank factor and x-ray quantum efficiency of a scintillator define the upper limit to its imaging performance, i.e. DQE(0). The Swank factor below the K-edge is dominated by optical properties, i.e. variations in light escape efficiency from different depths of interaction, denoted e(z). These variations can be optimized to improve tradeoffs in x-ray absorption, light yield, and spatial resolution. This work develops a quantitative model for interpreting measured PHS, and estimating e(z) on an absolute scale. The method is used to investigate segmented ceramic GOS scintillators used in multi-slice CT detectors. Methods: PHS of a ceramic GOS plate (1 mm thickness) and segmented GOS array (1.4 mm thick) were measured at 46 keV. Signal and noise propagation through x-ray conversion gain, light escape, detection by a photomultiplier tube and dynode amplification were modeled using a cascade of stochastic gain stages. PHS were calculated with these expressions and compared to measurements. Light escape parameters were varied until modeled PHS agreed with measurements. The resulting estimates of e(z) were used to calculate PHS without measurement noise to determine the inherent Swank factor. Results: The variation in e(z) was 67.2–89.7% in the plate and 40.2–70.8% in the segmented sample, corresponding to conversion gains of 28.6–38.1 keV{sup −1} and 17.1–30.1 keV{sup −1}, respectively. The inherent Swank factors of the plate and segmented sample were 0.99 and 0.95, respectively. Conclusion: The high light escape efficiency in the ceramic GOS samples yields high Swank factors and DQE(0) in CT applications. The PHS model allows the intrinsic optical properties of scintillators to be deduced from PHS measurements, thus it provides new insights for evaluating the imaging performance of segmented ceramic GOS scintillators.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Current-based metrology with two-terminal mesoscopic conductors

The traditional approach to quantum parameter estimation focuses on the quantum state, deriving fundamental bounds on precision through the quantum Fisher information. In most experimental settings, however, performing arbitrary quantum measurements is highly unfeasible. In open quantum systems, an alternative approach to metrology involves the measurement of stochastic currents flowing from the system to its environment. However, the present understanding of current-based metrology is mostly limited to Markovian master equations. Considering a parameter estimation problem in a two-terminal mesoscopic conductor, we identify the key elements that determine estimation precision within the Landauer-Büttiker formalism. Crucially, this approach allows us to address arbitrary coupling and temperature regimes. Furthermore, we obtain analytical results for the precision in linear-response and zero-temperature regimes. For the specific parameter estimation task that we consider, we demonstrate that the boxcar transmission function is optimal for current-based metrology in all parameter regimes.

Landauer formula↗

Catalyst: Fast and flexible modeling of reaction networks

We introduce Catalyst.jl, a flexible and feature-filled Julia library for modeling and high-performance simulation of chemical reaction networks (CRNs). Catalyst supports simulating stochastic chemical kinetics (jump process), chemical Langevin equation (stochastic differential equation), and reaction rate equation (ordinary differential equation) representations for CRNs. Through comprehensive benchmarks, we demonstrate that Catalyst simulation runtimes are often one to two orders of magnitude faster than other popular tools. More broadly, Catalyst acts as both a domain-specific language and an intermediate representation for symbolically encoding CRN models as Julia-native objects. This enables a pipeline of symbolically specifying, analyzing, and modifying CRNs; converting Catalyst models to symbolic representations of concrete mathematical models; and generating compiled code for numerical solvers. Leveraging ModelingToolkit.jl and Symbolics.jl, Catalyst models can be analyzed, simplified, and compiled into optimized representations for use in numerical solvers. Finally, we demonstrate Catalyst’s broad extensibility and composability by highlighting how it can compose with a variety of Julia libraries, and how existing open-source biological modeling projects have extended its intermediate representation.

59 BASIC BIOLOGICAL SCIENCES↗

Statistical Treatment of Inverse Problems Constrained by Differential Equations-based Models with Stochastic Terms

This paper introduces a statistical treatment of inverse problems constrained by models with stochastic terms. The solution of the forward problem is given by a distribution represented numerically by an ensemble of simulations. The goal is to formulate the inverse problem, in particular the objective function, to find the closest forward distribution (i.e., the output of the stochastic forward problem) that best explains the distribution of the observations in a certain metric. We use proper scoring rules, a concept employed in statistical forecast verification, namely energy, variogram, and hybrid (i.e., combination of the two) scores. We study the performance of the proposed formulation in the context of two applications: a coefficient field inversion for subsurface flow governed by an elliptic partial differential equation with a stochastic source and a parameter inversion for power grid governed by differential-algebraic equations. In both cases we show that the variogram and the hybrid scores produce better parameter inversion results than does the energy score, whereas the energy score leads to better probabilistic predictions.

97 MATHEMATICS AND COMPUTING↗

Bias-Variance Trade-Off in Physics-Informed Neural Networks with Randomized Smoothing for High-Dimensional PDEs

Physics-Informed Neural Networks (PINNs) have triggered a paradigm shift in scientific computing, leveraging mesh-free properties and robust approximation capabilities. While proving effective for low-dimensional partial differential equations (PDEs), the computational cost of PINNs remains a hurdle in high-dimensional scenarios. This is particularly pronounced when computing high-order and high-dimensional derivatives in the physics-informed loss. Randomized Smoothing PINN (RS-PINN) introduces Gaussian noise for stochastic smoothing of the original neural net model, enabling the use of Monte Carlo methods for derivative approximation, which eliminates the need for costly automatic differentiation. Despite its computational efficiency, especially in the approximation of high-dimensional derivatives, RS-PINN introduces biases in both loss and gradients, negatively impacting convergence, especially when coupled with stochastic gradient descent (SGD) algorithms. We present a comprehensive analysis of biases in RS-PINN, attributing them to the nonlinearity of the Mean Squared Error (MSE) loss as well as the intrinsic nonlinearity of the PDE itself. We propose tailored bias correction techniques, delineating their application based on the order of PDE nonlinearity. The derivation of an unbiased RS-PINN allows for a detailed examination of its advantages and disadvantages compared to the biased version. Specifically, the biased version has a lower variance and runs faster than the unbiased version, but it is less accurate due to the bias. To optimize the bias-variance trade-off, we combine the two approaches in a hybrid method that balances the rapid convergence of the biased version with the high accuracy of the unbiased version. In addition to methodological contributions, we present an enhanced implementation of RS-PINN. Extensive experiments on diverse high-dimensional PDEs, including Fokker-Planck, Hamilton-Jacobi-Bellman (HJB), viscous Burgers’, Allen-Cahn, and Sine-Gordon equations, illustrate the bias-variance trade-off and highlight the effectiveness of the hybrid RS-PINN. Empirical guidelines are provided for selecting biased, unbiased, or hybrid versions, depending on the dimensionality and nonlinearity of the specific PDE problem.

97 MATHEMATICS AND COMPUTING↗

The Discounted Method and Equivalence of Average Criteria for Risk-Sensitive Markov Decision Processes on Borel Spaces

This note concerns discrete-time controlled Markov chains with Borel state and action spaces. Given a nonnegative cost function, the performance of a control policy is measured by the superior limit risk-sensitive average criterion associated with a constant and positive risk sensitivity coefficient. Within such a framework, the discounted approach is used (a) to establish the existence of solutions for the corresponding optimality inequality, and (b) to show that, under mild conditions on the cost function, the optimal value functions corresponding to the superior and inferior limit average criteria coincide on a certain subset of the state space. The approach of the paper relies on standard dynamic programming ideas and on a simple analytical derivation of a Tauberian relation.

97 MATHEMATICS AND COMPUTING↗

A Copula-Based Conditional Probabilistic Forecast Model for Wind Power Ramps

Efficient management of wind ramping characteristics can significantly reduce wind integration costs for balancing authorities. By considering the stochastic dependence of wind power ramp (WPR) features, this paper develops a conditional probabilistic wind power ramp forecast (cp-WPRF) model based on Copula theory. The WPRs dataset is constructed by extracting ramps from a large dataset of historical wind power. Each WPR feature (e.g., rate, magnitude, duration, and start-time) is separately forecasted by considering the coupling effects among different ramp features. To accurately model the marginal distributions with a copula, a Gaussian mixture model (GMM) is adopted to characterize the WPR uncertainty and features. The Canonical Maximum Likelihood (CML) method is used to estimate parameters of the multivariable copula. The optimal copula model is chosen based on the Bayesian information criterion (BIC) from each copula family. Finally, the best conditions based cp-WPRF model is determined by predictive interval (PI) based evaluation metrics. Numerical simulations on publicly available wind power data show that the developed copula-based cp-WPRF model can predict WPRs with a high level of reliability and sharpness.

17 WIND ENERGY↗

A mixed $\ell_1$ regularization approach for sparse simultaneous approximation of parameterized PDEs

We introduce and assess a novel sparse polynomial technique for the simultaneous approximation of parameterized partial differential equations (PDEs) with deterministic and stochastic inputs. Our approach treats the numerical solution as a jointly sparse reconstruction problem through the reformulation of the standard basis pursuit denoising, where the set of jointly sparse vectors is infinite. To achieve global reconstruction of sparse solutions to parameterized elliptic PDEs over both physical and parametric domains, we combine the standard measurement scheme developed for compressed sensing in the context of bounded orthonormal systems with a novel mixed-norm based $\ell_1$ regularization method that exploits both energy and sparsity. Moreover, we are able to prove that, with minimal sample complexity, error estimates comparable to the best $s$-term and quasi-optimal approximations are achievable, while requiring only {\em a priori} bounds on polynomial truncation error with respect to the energy norm.Finally, we perform extensive numerical experiments on several high-dimensional parameterized elliptic PDE models to demonstrate the superior recovery properties of the proposed approach.

97 MATHEMATICS AND COMPUTING↗