Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “stochastic sampling”

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 307 records · Page 17
All sourcesNASA NTRSDOE OSTI
DOE OSTI2023

Sample-efficient verification of continuously-parameterized quantum gates for small quantum processors

Most near-term quantum information processing devices will not be capable of implementing quantum error correction and the associated logical quantum gate set. Instead, quantum circuits will be implemented directly using the physical native gate set of the device. These native gates often have a parameterization (e.g., rotation angles) which provide the ability to perform a continuous range of operations. Verification of the correct operation of these gates across the allowable range of parameters is important for gaining confidence in the reliability of these devices. In this work, we demonstrate a procedure for sample-efficient verification of continuously-parameterized quantum gates for small quantum processors of up to approximately 10 qubits. This procedure involves generating random sequences of randomly-parameterized layers of gates chosen from the native gate set of the device, and then stochastically compiling an approximate inverse to this sequence such that executing the full sequence on the device should leave the system near its initial state. We show that fidelity estimates made via this technique have a lower variance than fidelity estimates made via cross-entropy benchmarking. This provides an experimentally-relevant advantage in sample efficiency when estimating the fidelity loss to some desired precision. We describe the experimental realization of this technique using continuously-parameterized quantum gate sets on a trapped-ion quantum processor from Sandia QSCOUT and a superconducting quantum processor from IBM Q, and we demonstrate the sample efficiency advantage of this technique both numerically and experimentally.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2024

Microgrid design and multi-year dispatch optimization under climate-informed load and renewable resource uncertainty

Microgrids are an increasingly popular solution to provide energy resilience in response to increasing grid dependency and the growing impacts of climate change on grid operations. However, existing microgrid models do not currently consider the uncertain and long-term impacts of climate change when determining a set of design and operational decisions to minimize long-term costs or meet a resilience threshold. In this paper, we develop a novel scenario generation method that accounts for the uncertain effects of (i) climate change on variable renewable energy availability, (ii) extreme heat events on site load, and (iii) population and electrification trends on load growth. Additionally, we develop a two-stage stochastic programming extension of an existing microgrid design and dispatch optimization model to obtain uncertainty-informed and climate-resilient energy system decisions that minimizes long-term costs. Use of sample average approximation to validate our two case studies illustrates that the proposed methodology produces high-quality solutions that add resilience to systems with existing backup generation while reducing expected long-term costs.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2023

Blackout Diffusion: Generative Diffusion Models in Discrete-State Spaces

Typical generative diffusion models rely on a Gaussian diffusion process for training the backward transformations, which can then be used to generate samples from Gaussian noise. However, real world data often takes place in discrete-state spaces, including many scientific applications. Here, we develop a theoretical formulation for arbitrary discrete-state Markov processes in the forward diffusion process using exact (as opposed to variational) analysis. We relate the theory to the existing continuous-state Gaussian diffusion as well as other approaches to discrete diffusion, and identify the corresponding reverse-time stochastic process and score function in the continuous-time setting, and the reverse-time mapping in the discrete-time setting. As an example of this framework, we introduce “Blackout Diffusion”, which learns to produce samples from an empty image instead of from noise. Numerical experiments on the CIFAR-10, Binarized MNIST, and CelebA datasets confirm the feasibility of our approach. Generalizing from specific (Gaussian) forward processes to discrete-state processes without a variational approximation sheds light on how to interpret diffusion models, which we discuss.

Santos, Javier E.↗
NASA NTRS2003

Stochastic Mixing Model with Power Law Decay of Variance

Here we present a simple stochastic mixing model based on the law of large numbers (LLN). The reason why the LLN is involved in our formulation of the mixing problem is that the random conserved scalar c = c(t,x(t)) appears to behave as a sample mean. It converges to the mean value mu, while the variance sigma(sup 2)(sub c) (t) decays approximately as t(exp -1). Since the variance of the scalar decays faster than a sample mean (typically is greater than unity), we will introduce some non-linear modifications into the corresponding pdf-equation. The main idea is to develop a robust model which is independent from restrictive assumptions about the shape of the pdf. The remainder of this paper is organized as follows. In Section 2 we derive the integral equation from a stochastic difference equation describing the evolution of the pdf of a passive scalar in time. The stochastic difference equation introduces an exchange rate gamma(sub n) which we model in a first step as a deterministic function. In a second step, we generalize gamma(sub n) as a stochastic variable taking fluctuations in the inhomogeneous environment into account. In Section 3 we solve the non-linear integral equation numerically and analyze the influence of the different parameters on the decay rate. The paper finishes with a conclusion.

Fedotov, S.↗
NASA NTRS1976

Analysis of bilinear noise models in circuits and devices

There are a number of applications in which linear noise models are inappropriate. In the paper, the use of bilinear noise models in circuits and devices is considered. Several physical problems are studied in this framework. These include circuits involving varying parameters (such as variable resistance circuits constructed using field-effect transistors), the effect of switching jitter on sampled data system performance and communication systems involving voltage-controlled oscillators and phase-lock loops. In addition, several types of analytical techniques for stochastic bilinear systems are considered. Specifically, the moment equations of Brockett for bilinear systems driven by white noise are discussed, and closed-form expressions for certain bilinear systems (those that evolve an Abelian or solvable Lie groups) driven by white or colored noise are derived. In addition, an approximate statistical technique involving the use of harmonic expansions is described.

Willsky, A. S.↗
NASA NTRS1983

Lagged average forecasting, an alternative to Monte Carlo forecasting

A 'lagged average forecast' (LAF) model is developed for stochastic dynamic weather forecasting and used for predictions in comparison with the results of a Monte Carlo forecast (MCF). The technique involves the calculation of sample statistics from an ensemble of forecasts, with each ensemble member being an ordinary dynamical forecast (ODF). Initial conditions at a time lagging the start of the forecast period are used, with varying amounts of time for the lags. Forcing by asymmetric Newtonian heating of the lower layer is used in a two-layer, f-plane, highly truncated spectral model in a test forecasting run. Both the LAF and MCF are found to be more accurate than the ODF due to ensemble averaging with the MCF and the LAF. When a regression filter is introduced, all models become more accurate, with the LAF model giving the best results. The possibility of generating monthly or seasonal forecasts with the LAF is discussed.

Hoffman, R. N.↗
NASA NTRS1987

Combined optimization of image-gathering and image-processing systems for scene feature detection

The relationship between the image gathering and image processing systems for minimum mean squared error estimation of scene characteristics is investigated. A stochastic optimization problem is formulated where the objective is to determine a spatial characteristic of the scene rather than a feature of the already blurred, sampled and noisy image data. An analytical solution for the optimal characteristic image processor is developed. The Wiener filter for the sampled image case is obtained as a special case, where the desired characteristic is scene restoration. Optimal edge detection is investigated using the Laplacian operator x G as the desired characteristic, where G is a two dimensional Gaussian distribution function. It is shown that the optimal edge detector compensates for the blurring introduced by the image gathering optics, and notably, that it is not circularly symmetric. The lack of circular symmetry is largely due to the geometric effects of the sampling lattice used in image acquisition. The optimal image gathering optical transfer function is also investigated and the results of a sensitivity analysis are shown.

Halyo, Nesim↗
DOE OSTI2023

ClimSim: A large multi-scale dataset for hybrid physics-ML climate emulation

Modern climate projections lack adequate spatial and temporal resolution due to computational constraints. A consequence is inaccurate and imprecise predictions of critical processes such as storms. Hybrid methods that combine physics with machine learning (ML) have introduced a new generation of higher fidelity climate simulators that can sidestep Moore’s Law by outsourcing compute-hungry, short, high-resolution simulations to ML emulators. However, this hybrid ML-physics simulation approach requires domain-specific treatment and has been inaccessible to MLexperts because of lack of training data and relevant, easy-to-use workflows. Wepresent ClimSim, the largest-ever dataset designed for hybrid ML-physics research. It comprises multi-scale climate simulations, developed by a consortium of climate scientists and ML researchers. It consists of 5.7 billion pairs of multivariate input and output vectors that isolate the influence of locally-nested, high-resolution, high-fidelity physics on a host climate simulator’s macro-scale physical state. The dataset is global in coverage, spans multiple years at high sampling frequency, and is designed such that resulting emulators are compatible with downstream coupling into operational climate simulators. We implement a range of deterministic and stochastic regression baselines to highlight the ML challenges and their scoring. The data (https://huggingface.co/datasets/LEAP/ClimSim_high-res2) and code(https://leap-stc.github.io/ClimSim)arereleasedopenlytosupport the development of hybrid ML-physics and high-fidelity climate simulations for the benefit of science and society.

artificial intelligence, machine learning↗
DOE OSTI2020

Real-time Ridesharing for Transportation Hubs with Demand and Supply Uncertainty

Transportation hubs in major cities generate a significant amount of trips by taxis and for-hail vehicles (FHV), with many of the trips sharing similar destinations. This suggests promising opportunities to leverage the collective travel needs with dedicated ridesharing solutions to reduce the externalities of excessive traffic at transportation hubs. In this study, we develop a novel dynamic ridesharing approach to serve trips from the transportation hub by considering (1) demand (new passengers) and supply (newly available vehicles) in the near future and (2) the uncertainty of future predictions. Our approach consists of two stages. In the first stage, we develop a data structure called hub mobility tree to generate potential combinations of shareable trips as candidate schedules efficiently. Then the generated schedules are used in the second stage to formulate the stochastic hub-based ridesharing problem (SHRP), which is a stochastic integer programming problem with the objective to maximize the total expected ridesharing profit over time. Due to the prohibitive number of shareable trips, we then approximately solve SHRP by the sample average approximate method (SAA), and a dual Lagrangian technique is implemented to further improve the scalability of the solution approach. We demonstrate the performances of the proposed method by simulating the ridesharing service at JFK airport using NYC taxi and FHV data. The results indicate that the proposed method outperforms the myopic ridesharing (maximize profit for a single time step) and the rolling horizon method with point estimation of future demand and supply.

dynamic ridesharing↗
NASA NTRS2002

Scale Dependence of Statistics of Spatially Averaged Rain Rate Seen in TOGA COARE Comparison with Predictions from a Stochastic Model

A characteristic feature of rainfall statistics is that they in general depend on the space and time scales over which rain data are averaged. As a part of an earlier effort to determine the sampling error of satellite rain averages, a space-time model of rainfall statistics was developed to describe the statistics of gridded rain observed in GATE. The model allows one to compute the second moment statistics of space- and time-averaged rain rate which can be fitted to satellite or rain gauge data to determine the four model parameters appearing in the precipitation spectrum - an overall strength parameter, a characteristic length separating the long and short wavelength regimes and a characteristic relaxation time for decay of the autocorrelation of the instantaneous local rain rate and a certain 'fractal' power law exponent. For area-averaged instantaneous rain rate, this exponent governs the power law dependence of these statistics on the averaging length scale $L$ predicted by the model in the limit of small $L$. In particular, the variance of rain rate averaged over an $L \times L$ area exhibits a power law singularity as $L \rightarrow 0$. In the present work the model is used to investigate how the statistics of area-averaged rain rate over the tropical Western Pacific measured with ship borne radar during TOGA COARE (Tropical Ocean Global Atmosphere Coupled Ocean Atmospheric Response Experiment) and gridded on a 2 km grid depends on the size of the spatial averaging scale. Good agreement is found between the data and predictions from the model over a wide range of averaging length scales.

Kundu, Prasun K.↗
NASA NTRS2010

Analyzing Non Stationary Processes in Radiometers

The lack of well-developed techniques for modeling changing statistical moments in our observations has stymied the application of stochastic process theory for many scientific and engineering applications. Non linear effects of the observation methodology is one of the most perplexing aspects to modeling non stationary processes. This perplexing problem was encountered when modeling the effect of non stationary receiver fluctuations on the performance of radiometer calibration architectures. Existing modeling approaches were found not applicable; particularly problematic is modeling processes across scales over which they begin to exhibit non stationary behavior within the time interval of the calibration algorithm. Alternatively, the radiometer output is modeled as samples from a sequence random variables; the random variables are treated using a conditional probability distribution function conditioned on the use of the variable in the calibration algorithm. This approach of treating a process as a sequence of random variables with non stationary stochastic moments produce sensible predictions of temporal effects of calibration algorithms. To test these model predictions, an experiment using the Millimeter wave Imaging Radiometer (MIR) was conducted. The MIR with its two black body calibration references was configured in a laboratory setting to observe a third ultra-stable reference (CryoTarget). The MIR was programmed to sequentially sample each of the three references in approximately a 1 second cycle. Data were collected over a six-hour interval. The sequence of reference measurements form an ensemble sample set comprised of a series of three reference measurements. Two references are required to estimate the receiver response. A third reference is used to estimate the uncertainty in the estimate. Typically, calibration algorithms are designed to suppress the non stationary effects of receiver fluctuations. By treating the data sequence as an ensemble collection, it is possible to apply temporal algorithms which exacerbate the non stationary effects. By varying the algorithm, information about the properties of the non stationary receiver fluctuations is obtained. Comparisons of analytical calculations and statistical analysis of data demonstrate impressive agreement.

Racette, Paul↗
NASA NTRS2010

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.↗
DOE OSTI2023

Exploiting stochastic locality in lattice QCD: hadronic observables and their uncertainties

Abstract Because of the mass gap, lattice QCD simulations exhibit stochastic locality: distant regions of the lattice fluctuate independently. There is a long history of exploiting this to increase statistics by obtaining multiple spatially-separated samples from each gauge field; in the extreme case, we arrive at the master-field approach in which a single gauge field is used. Here we develop techniques for studying hadronic observables using position-space correlators, which are more localized, and compare with the standard time-momentum representation. We also adapt methods for estimating the variance of an observable from autocorrelated Monte Carlo samples to the case of correlated spatially-separated samples.

Physics↗
DOE OSTI2020

Benchmark Solutions for Radiation Transport in Stochastic Media with Inhomogeneous Material Statistics

Accurately solving implicit Monte Carlo (IMC) thermal photon transport problems with mixed material cells is important in realistic applications. The production IMC package at LLNL treats mixed material cells arising from ALE remap and hydrodynamics using the same approximate model. The new Imp IMC thermal photon transport package currently under development has both a material interface reconstruction (MIR) algorithm and a Levermore-Pomraning (LP) stochastic medium algorithm for treating mixed material cells. Existing stochastic medium algorithms for treating mixed material cells in IMC lack a complete theoretical basis. The IMC LP algorithm implementation has been demonstrated to reproduce published deterministic LP solutions for the particular case of spatially homogeneous material statistics. Realistic simulations will include spatially inhomogeneous material statistics (material mean chord lengths). In a previous investigation, the LP-model for transport in binary stochastic media in rod geometry was generalized to accommodate spatially varying material chord lengths, i.e., the mixing statistics were allowed to be nonhomogeneous. Analytical solutions were obtained and used to produce a verifi cation suite for the Imp IMC Levermore-Pomraning implementation for different spatial variations of the chord lengths. However, the accuracy of the LP model when the mixing statistics are nonhomogeneous has not been assessed and leaves open the question of whether local accuracy is improved or further degraded when chord lengths are not uniform. This shortcoming is rectifi ed here by developing benchmark analytic solutions for transport in binary Markovian stochastic mixtures in rod geometry with nonhomogeneous mixing statistics, using spatially varying chord lengths considered in the previous investigation based on the LP model. Methods for sampling a nonhomogeneous Poisson process (NHPP) are first described and used to construct individual realizations of the binary mixtures in rod geometry. Analytic solutions are then obtained for the forward and backward directed fluxes on a given realization, now viewed as a deterministic medium with alternating layers of the two materials with known interface locations. Finally, material averaged scalar fluxes are obtained using these sampling schemes with spatially linear and quadratic chord lengths and used to assess the accuracy of the previously obtained LP-model results.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
NASA NTRS1986

Combined optimization of image-gathering optics and image-processing algorithm for edge detection

This paper investigates the relationships between the image-gathering and image-processing systems for minimum mean-squared error estimation of scene characteristics. A stochastic optimization problem is formulated in which the objective is to determine a spatial characteristic of the scene rather than a feature of the already blurred, sampled, and noisy image data. The Wiener filter for the sampled image case is obtained as a special case, where the desired characteristics is scene restoration. Optimal edge detection is investigated. It is shown that the optimal edge detector compensates for the blurring introduced by the image-gathering optics, and, notably, that it is not circularly symmetric. The lack of circular symmetry is largely due to the geometric effects of the sampling lattice used in image acquisition.

Halyo, N.↗
NASA NTRS1989

Electron acceleration by an obliquely propagating electromagnetic wave in the regime of validity of the Fokker-Planck-Kolmogorov approach

The relativistic motion of an ensemble of electrons in an intense monochromatic electromagnetic wave propagating obliquely in a uniform external magnetic field is studied. The problem is formulated from the viewpoint of Hamiltonian theory and the Fokker-Planck-Kolmogorov approach analyzed by Hizanidis (1989), leading to a one-dimensional diffusive acceleration along paths of constant zeroth-order generalized Hamiltonian. For values of the wave amplitude and the propagating angle inside the analytically predicted stochastic region, the numerical results suggest that the diffusion probes proceeds in stages. In the first stage, the electrons are accelerated to relatively high energies by sampling the first few overlapping resonances one by one. During that stage, the ensemble-average square deviation of the variable involved scales quadratically with time. During the second stage, they scale linearly with time. For much longer times, deviation from linear scaling slowly sets in.

Hizanidis, Kyriakos↗
DOE OSTI2021

Adversarial sampling of unknown and high-dimensional conditional distributions

Many engineering problems require the prediction of realization-to-realization variability or a refined description of modeled quantities. In that case, it is necessary to sample elements from unknown high-dimensional spaces with possibly millions of degrees of freedom. While there exist methods able to sample elements from probability density functions (PDF) with known shapes, several approximations need to be made when the distribution is unknown. In this paper the sampling method, as well as the inference of the underlying distribution, are both handled with a data-driven method known as generative adversarial networks (GAN), which trains two competing neural networks to produce a network that can effectively generate samples from the training set distribution. In practice, it is often necessary to draw samples from conditional distributions. When the conditional variables are continuous, only one (if any) data point corresponding to a particular value of a conditioning variable may be available, which is not sufficient to estimate the conditional distribution. This work handles this problem using an a priori estimation of the conditional moments of a PDF. Herein, two approaches, stochastic estimation, and an external neural network are compared for computing these moments; however, any preferred method can be used. The algorithm is demonstrated in the case of the deconvolution of a filtered turbulent flow field. It is shown that all the versions of the proposed algorithm effectively sample the target conditional distribution with minimal impact on the quality of the samples compared to state-of-the-art methods. Additionally, the procedure can be used as a metric for the diversity of samples generated by a conditional GAN (cGAN) conditioned with continuous variables.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2025

Score-Based Physics-Informed Neural Networks for High-Dimensional Fokker–Planck Equations

The Fokker-Planck (FP) equation is a foundational partial differential equation (PDE) in stochastic processes involving Brownian motions. However, the curse of dimensionality (CoD) poses a formidable challenge when dealing with high-dimensional FP equations. Although Monte Carlo simulation and (vanilla) Physics-Informed Neural Networks (PINNs) have shown the potential to tackle CoD, both methods exhibit significant numerical errors in high dimensions when dealing with the probability density function (PDF) associated with Brownian motion. The point-wise PDF values tend to decrease exponentially as dimensionality increases, surpassing the precision of numerical simulations and resulting in substantial errors. In addition, due to its massive sampling, Monte Carlo fails to offer fast sampling. Modeling the logarithm likelihood (LL) via vanilla PINNs transforms the FP equation into a notoriously difficult Hamilton-Jacobi-Bellman (HJB) equation, which is impractical for PINN learning, whose error grows rapidly with dimension. To this end, we propose a novel approach utilizing a score-based solver to fit the score function in stochastic differential equations (SDEs). The score function, defined as the gradient of the LL, plays a fundamental role in inferring LL and PDF and enables fast SDE sampling, offering an effective means to overcome the CoD. Three fitting methods, Score Matching (SM), Sliced Score Matching (SSM), and Score-PINN, are introduced, each contributing unique advantages in computational complexity, accuracy, and generality. The proposed score-based SDE solver operates in two stages: first, employing score matching or Score-PINN to acquire the score function; and second, solving the LL via an ordinary differential equation (ODE) using the obtained score function. Comparative evaluations across these methods showcase varying trade-offs. The proposed methodology is evaluated across diverse SDEs, including anisotropic Ornstein-Uhlenbeck processes, geometric Brownian motion, and Brownian motion with varying eigenspace. We also test various distributions, including Gaussian, Log-normal, Laplace, and Cauchy distributions. The numerical results demonstrate the score-based SDE solver’s stability, speed, and performance across different experimental settings, solidifying its potential as a solution to CoD for high-dimensional FP equations.

97 MATHEMATICS AND COMPUTING↗