Engineering Papers⌕ Search

SEARCH · Engineering Papers

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 973 records · Page 54

Accuracy of Some Approximate Gaussian Filters for the Navier--Stokes Equation in the Presence of Model Error

Bayesian state estimation of a dynamical system from a stream of noisy measurements is important in many geophysical and engineering applications where high dimensionality of the state space, sparse observations, and model error pose key challenges. In this work, three computationally feasible, approximate Gaussian data assimilation/filtering algorithms are considered in various regimes of turbulent 2D Navier--Stokes dynamics in the presence of model error. The first source of error arises from the necessary use of reduced models for the forward dynamics of the filters, while a particular type of representation error arises from the finite resolution of observations which mix up information about resolved and unresolved dynamics. Two stochastically parameterized filtering algorithms, referred to as cSPEKF and GCF, are compared with 3DVAR---a prototypical time-sequential algorithm known to be accurate for filtering dissipative systems for a suitably inflated “background” covariance. We provide the first evidence that the stochastically parameterized algorithms, which do not rely on detailed knowledge of the underlying dynamics and do not require covariance inflation, can compete with or outperform an optimally tuned 3DVAR algorithm, and they can overcome competing sources of error in a range of dynamical scenarios.

97 MATHEMATICS AND COMPUTING↗

Optimal Complementarity Analysis of Potential Floating Solar Co-Located With Existing Hydropower Assets Across the Contiguous United States

The U.S. is expected to double its rate of renewable capacity from 2024 to 2030. However, the stochastic nature of renewable energy poses challenges to the operation and reliability of our power grid. The combined generation from renewable energy sources, with dispatchable sources (such as hydropower) operating as a hybrid energy plant, could mitigate this variability. In this paper, the complementarity analysis of selected U.S. reservoirs with existing hydropower assets (EHAs) and potential floating photovoltaics (FPVs) is conducted for the continuous U.S. (CONUS). The optimal FPV capacity for each site is determined by minimizing the variability of the combined output, while adhering to the FPV potential. Our results indicate that over 50% of the analyzed reservoirs achieve a stability coefficient exceeding 0.5, leading to a less-variable output after optimization. Finally, we analyze the complementary hydro-FPV hybrid reservoirs by considering both the Pearson correlation coefficient and the stability coefficient on daily, monthly, and yearly scales. Summaries are included of locations of theoretical FPVs co-located with hydropower plants that exhibit high complementarity based on the selected metrics.

13 - HYDRO ENERGY↗

Optimization of rotational arc station parameter optimized radiation therapy

Purpose: To develop a fast optimization method for station parameter optimized radiation therapy (SPORT) and show that SPORT is capable of matching VMAT in both plan quality and delivery efficiency by using three clinical cases of different disease sites. Methods: The angular space from 0° to 360° was divided into 180 station points (SPs). A candidate aperture was assigned to each of the SPs based on the calculation results using a column generation algorithm. The weights of the apertures were then obtained by optimizing the objective function using a state-of-the-art GPU based proximal operator graph solver. To avoid being trapped in a local minimum in beamlet-based aperture selection using the gradient descent algorithm, a stochastic gradient descent was employed here. Apertures with zero or low weight were thrown out. To find out whether there was room to further improve the plan by adding more apertures or SPs, the authors repeated the above procedure with consideration of the existing dose distribution from the last iteration. At the end of the second iteration, the weights of all the apertures were reoptimized, including those of the first iteration. The above procedure was repeated until the plan could not be improved any further. The optimization technique was assessed by using three clinical cases (prostate, head and neck, and brain) with the results compared to that obtained using conventional VMAT in terms of dosimetric properties, treatment time, and total MU. Results: Marked dosimetric quality improvement was demonstrated in the SPORT plans for all three studied cases. For the prostate case, the volume of the 50% prescription dose was decreased by 22% for the rectum and 6% for the bladder. For the head and neck case, SPORT improved the mean dose for the left and right parotids by 15% each. The maximum dose was lowered from 72.7 to 71.7 Gy for the mandible, and from 30.7 to 27.3 Gy for the spinal cord. The mean dose for the pharynx and larynx was reduced by 8% and 6%, respectively. For the brain case, the doses to the eyes, chiasm, and inner ears were all improved. SPORT shortened the treatment time by ∼1 min for the prostate case, ∼0.5 min for brain case, and ∼0.2 min for the head and neck case. Conclusions: The dosimetric quality and delivery efficiency presented here indicate that SPORT is an intriguing alternative treatment modality. With the widespread adoption of digital linac, SPORT should lead to improved patient care in the future.

60 APPLIED LIFE SCIENCES↗

Wireless Fading Channel Models: From Classical to Stochastic Differential Equations

The wireless communications channel constitutes the basic physical link between the transmitter and the receiver antennas. Its modeling has been and continues to be a tantalizing issue, while being one of the most fundamental components based on which transmitters and receivers are designed and optimized. The ultimate performance limits of any communication system are determined by the channel it operates in. Realistic channel models are thus of utmost importance for system design and testing. In addition to exponential power path-loss, wireless channels suffer from stochastic short term fading (STF) due to multipath, and stochastic long term fading (LTF) due to shadowing depending on the geographical area. STF corresponds to severe signal envelope fluctuations, and occurs in densely built-up areas filled with lots of objects like buildings, vehicles, etc. On the other hand, LTF corresponds to less severe mean signal envelope fluctuations, and occurs in sparsely populated or suburban areas. In general, LTF and STF are considered as superimposed and may be treated separately. Ossanna was the pioneer to characterize the statistical properties of the signal received by a mobile user, in terms of interference of incident and reflected waves. His model was better suited for describing fading occurring mainly in suburban areas (LTF environments). It is described by the average power loss due to distance and power loss due to reflection of signals from surfaces, which when measured in dB's give rise to normal distributions, and this implies that the channel attenuation coefficient is log-normally distributed. Furthermore, in mobile communications, the LTF channel models are also characterized by their special correlation characteristics which have been reported. Clarke introduced the first comprehensive scattering model describing STF occurring mainly in urban areas. An easy way to simulate Clarke's model using a computer simulation is described. This model was later expanded to three-dimensions (3D) by Aulin. An indoor STF was introduced. Most of these STF models provide information on the frequency response of the channel, described by the Doppler power spectral density (DPSD). Aulin presented a methodology to find the Doppler power spectrum by computing the Fourier transform of the autocorrelation function of the channel impulse response with respect to time. A different approach, leading to the same Doppler power spectrum relation was presented by Gans. These STF models suggest various distributions for the received signal amplitude such as Rayleigh, Rician, or Nakagami. Models based on autoregressive and moving averages (AR) are proposed. However, these models assume that the channel state is completely observable, which in reality is not the case due to additive noise, and requires long observation intervals. First order Markov models for Raleigh fading have been proposed, and the usefulness of a finite-state Markov channel model is argued. Mobile-to-mobile (or ad hoc) wireless networks comprise nodes that freely and dynamically self-organize into arbitrary and/or temporary network topology without any fixed infrastructure support. They require direct communication between a mobile transmitter and a mobile receiver over a wireless medium. Such mobile-to-mobile communication systems differ from the conventional cellular systems, where one terminal, the base station, is stationary, and only the mobile station is moving. As a consequence, the statistical properties of mobile-to-mobile links are different from cellular ones. Copious ad hoc networking research exists on layers in the open system interconnection (OSI) model above the physical layer. However, neglecting the physical layer while modeling wireless environment is error prone and should be considered more carefully. The experimental results show that the factors at the physical layer not only affect the absolute performance of a protocol, but because their impact on different protocols is nonuniform, it can even change the relative ranking among protocols for the same scenario. The importance of the physical layer is demonstrated by evaluating the Medium Access Control (MAC) performance. Most of the research conducted on wireless channel modeling deals mainly with deterministic wireless channel models. In these models, the speeds of the nodes are assumed to be constant and the statistical characteristics of the received signal are assumed to be fixed with time. But in reality, the propagation environment varies continuously due to mobility of the nodes at variable speeds and movement of objects or scatter across transmitters and receivers resulting in appearance or disappearance of existing paths from one instant to the next. As a result, the current models that assume fixed statistics are unable to capture and track complex time variations in the propagation environment.

99 GENERAL AND MISCELLANEOUS↗

Revisiting the ODE Method for Recursive Algorithms: Fast Convergence Using Quasi Stochastic Approximation

Several decades ago, Profs. Sean Meyn and Lei Guo were postdoctoral fellows at ANU, where they shared interest in recursive algorithms. It seems fitting to celebrate Lei Guo's 60th birthday with a review of the ODE Method and its recent evolution. The method has been regarded as a technique for algorithm analysis. It is argued that this viewpoint is backwards: The original stochastic approximation method was surely motivated by an ODE, and tools for analysis came much later (based on establishing robustness of Euler approximations). The paper presents a brief survey of recent research in machine learning that shows the power of algorithm design in continuous time, following by careful approximation to obtain a practical recursive algorithm. While these methods are usually presented in a stochastic setting, this is not a prerequisite. In fact, recent theory shows that rates of convergence can be dramatically accelerated by applying techniques inspired by quasi Monte-Carlo. Subject to conditions, the optimal rate of convergence can be obtained by applying the averaging technique of Polyak and Ruppert. The conditions are not universal, but theory suggests alternatives to achieve acceleration. The theory is illustrated with applications to gradient-free optimization, and policy gradient algorithms for reinforcement learning.

learning and adaptive systems in artificial intell↗

Reactive power planning under high penetration of wind energy using Benders decomposition

This study addresses the optimal allocation of reactive power volt-ampere reactive (VAR) sources under the paradigm of high penetration of wind energy. Reactive power planning (RPP) in this particular condition involves a high level of uncertainty because of wind power characteristic. To properly model wind generation uncertainty, a multi-scenario framework optimal power flow that considers the voltage stability constraint under the worst wind scenario and transmission N 1 contingency is developed. The objective of RPP in this study is to minimise the total cost including the VAR investment cost and the expected generation cost. Therefore RPP under this condition is modelled as a two-stage stochastic programming problem to optimise the VAR location and size in one stage, then to minimise the fuel cost in the other stage, and eventually, to find the global optimal RPP results iteratively. Benders decomposition is used to solve this model with an upper level problem (master problem) for VAR allocation optimisation and a lower problem (sub-problem) for generation cost minimisation. Impact of the potential reactive power support from doubly-fed induction generator (DFIG) is also analysed. Lastly, case studies on the IEEE 14-bus and 118-bus systems are provided to verify the proposed method.

17 WIND ENERGY↗

Folding of small knotted proteins: Insights from a mean field coarse-grained model

A small but relevant number of proteins whose native structure is known features nontrivial topology, i.e., they are knotted. Understanding the process of folding from a swollen unknotted state to the biologically relevant native conformation is, for these proteins, particularly difficult, due to their rate-limiting topological entanglement. To shed some light into this conundrum, we introduced a structure-based coarse-grained model of the protein, where the information about the folded conformation is encoded in bonded angular interactions only, which do not favor the formation of native contacts. A stochastic search scheme in parameter space is employed to identify a set of interactions that maximizes the probability to attain the knotted state. The optimal knotting pathways of the two smallest knotted proteins, obtained through this approach, are consistent with the results derived by means of coarse-grained as well as full atomistic simulations.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Integrated path planning and control through proximal policy optimization for a marine current turbine

This paper presents an integrated path planning and tracking control framework for a marine current turbine (MCT), where the MCT is treated as an energy-harvesting autonomous underwater vehicle (AUV). Considering the ocean (space of action) is continuous, the proposed framework employs two modules to address path planning and path tracking enabled by the proximal policy optimization (PPO) algorithm, which is a policy gradient deep reinforcement learning (RL) method. Further, to enable fully autonomous operation in a stochastic oceanic environment, the proposed path planning seeks a primary objective of maximizing the harvested energy; then, the path tracking module is designed to minimize the tracking error and avoid collisions with static and dynamic obstacles. Using field-collected acoustic Doppler current profiler (ADCP) data, the performance of the proposed framework is evaluated. Comparative studies with baseline algorithms in three different scenarios of path planning, path tracking without an obstacle, and path tracking with collision avoidance verify the effectiveness of our proposed approach.

16 TIDAL AND WAVE POWER↗

Stochastic minibatch approach to the ptychographic iterative engine

The ptychographic iterative engine (PIE) is a widely used algorithm that enables phase retrieval at nanometer-scale resolution over a wide range of imaging experiment configurations. By analyzing diffraction intensities from multiple scanning locations where a probing wavefield interacts with a sample, the algorithm solves a difficult optimization problem with constraints derived from the experimental geometry as well as sample properties. The effectiveness at which this optimization problem is solved is highly dependent on the ordering in which we use the measured diffraction intensities in the algorithm, and random ordering is widely used due to the limited ability to escape from stagnation in poor-quality local solutions. In this study, we introduce an extension to the PIE algorithm that uses ideas popularized in recent machine learning training methods, in this case minibatch stochastic gradient descent. Our results demonstrate that these new techniques significantly improve the convergence properties of the PIE numerical optimization problem.

47 OTHER INSTRUMENTATION↗

Physics-based statistical learning approach to mesoscopic model selection

In materials science and many other research areas, models are frequently inferred without considering their generalization to unseen data. Here, we apply statistical learning using cross-validation to obtain an optimally predictive coarse-grained description of a two-dimensional kinetic nearest-neighbor Ising model with Glauber dynamics (GD) based on the stochastic Ginzburg-Landau equation (sGLE). The latter is learned from GD “training” data using a log-likelihood analysis, and its predictive ability for various complexities of the model is tested on GD “test” data independent of the data used to train the model on. Using two different error metrics, we perform a detailed analysis of the error between magnetization time trajectories simulated using the learned sGLE coarse-grained description and those obtained using the GD model. We show that both for equilibrium and out-of-equilibrium GD training trajectories, the standard phenomenological description using a quartic free energy does not always yield the most predictive coarse-grained model. Moreover, increasing the amount of training data can shift the optimal model complexity to higher values. Our results are promising in that they pave the way for the use of statistical learning as a general tool for materials modeling and discovery.

36 MATERIALS SCIENCE↗

Sustainable network design for multi-purpose pellet processing depots under biomass supply uncertainty

This work develops a two-stage stochastic mixed-integer programming model to manage multi- purpose pellet processing depots under feedstock supply uncertainty. The proposed optimization model would help to minimize cost and to mitigate emissions from the supply chain network. We consider three alternative Biomass Processing and Densi cation Depot (BPDD) technologies; namely, conventional pellet processing, high moisture pellet processing, and ammonia ber expansion. These three technologies pre-process/pre-treat and densify different types of biomass into more highly densi ed intermediate products for different markets in order to improve movability and overall supply network performance in terms of costs and emissions. A hybrid decomposition algorithm was developed that combines sample average approximation with an enhanced Progressive Hedging (PH) algorithm to solve this challenging NP -hard problem. Some heuristics such as Rolling Horizon (RH) heuristic, variable xing technique were later applied to further enhance the PH algorithm. Mississippi and Alabama were selected as a testing ground and ArcGIS was employed to visualize and validate the modeling results. The results of the analysis reveal promising insights that could lead to recommendations to help decision makers achieve a more cost-effective environmentally-friendly supply chain network.

09 BIOMASS FUELS↗

Enabling scientific machine learning in MOOSE using Libtorch

A neural-network-based machine learning interface has been developed for the Multiphysics Object-Oriented Simulation Environment (MOOSE). The interface relies on Libtorch, the C++ front-end of PyTorch, and enables an online interaction between modern machine learning algorithms and all the existing simulation, modeling, and analysis processes available in MOOSE. New capabilities in MOOSE include the native generation and training of artificial neural networks together with options to load pretrained neural networks in TorchScript format. Furthermore, the MOOSE stochastic tools module (MOOSE-STM) has been enhanced with neural network-based surrogate and reduced-order model generation options for efficient stochastic analyses. Lastly, a reinforcement learning capability has been added to MOOSE-STM for the interactive control and optimization of complex multiphysics problems.

97 MATHEMATICS AND COMPUTING↗

A Logarithmic Bayesian Approach to Quantum Error Detection

We consider the problem of continuous quantum error correction from a Bayesian perspective, proposing a pair of digital filters using logarithmic probabilities that are able to achieve near-optimal performance on a three-qubit bit-flip code, while still being reasonable to implement on low-latency hardware. These practical filters are approximations of an optimal filter that we derive explicitly for finite time steps, in contrast with previous work that has relied on stochastic differential equations such as the Wonham filter. By utilizing logarithmic probabilities, we are able to eliminate the need for explicit normalization and can reduce the Gaussian noise distribution to a simple quadratic expression. The state transitions induced by the bit-flip errors are modeled using a Markov chain, which for log-probabilities must be evaluated using a LogSumExp function. We develop the two versions of our filter by constraining this LogSumExp to have either one or two inputs, which favors either simplicity or accuracy, respectively. Using simulated data, we demonstrate that the single-term and two-term filters are able to significantly outperform both a double threshold scheme and a linearized version of the Wonham filter in tests of error detection under a wide variety of error rates and time steps.

97 MATHEMATICS AND COMPUTING↗

Implementation of longitudinal dynamics with barrier RF in BETACOOL and comparison to ESME

The barrier bucket RF system is successfully used on Recycler storage ring at Fermilab. The special program code ESME was used for numerical simulation of longitudinal phase space manipulations. This program helps optimizing the various regimes of operation in the Recycler and increasing the luminosity in the colliding experiments. Electron and stochastic cooling increases the phase space density in all degrees of freedom. In the case of a small phase space volume the intrabeam scattering introduces coupling between the transverse and longitudinal temperatures of the antiproton beam. For numerical simulations of the cooling processes at the Recycler, a new model of the barrier buckets was implemented in the BETACOOL code. The comparison between ESME and BETACOOL codes for a stationary and moving barrier buckets is presented. This article also includes an application of the barrier bucket numerical model for simulation of the luminosity distribution for RHIC colliding experiments. These simulations take into account the specific longitudinal distribution of the bunch and the vertex size of the detector.

43 PARTICLE ACCELERATORS↗

PDE-constrained Optimization under Uncertainty

Uncertainty is pervasive in all science and engineering applications. Incorporating uncertainty in physical models is therefore both natural and vital. In doing so, we often arrive at parametric systems of partial differential equations (PDEs). When passing from simulation to optimization, we obtain (typically nonconvex) infinite-dimensional optimization problems that, upon discretization, result in extremely large-scale nonlinear programs. For example, consider a linear elliptic PDE on a two- dimensional domain with a single random coeffcient. If we sampled the random input with 10,000 realizations of the coeffcient, the resulting optimization problem would have 10,000 PDE constraints. Furthermore, discretizing each PDE with piecewise linear finite elements on a 100X100 uni- form quadrilateral mesh results in 100,000,000 degrees of freedom. As a result, the critical components for ensuring mesh-independent performance of numerical optimization methods in the deterministic setting, for example, solution regularity and generalized differentiability, are even more critical in the stochastic setting.

42 ENGINEERING↗

Particle Swarm Optimisation for group structure optimization for radiotherapy shielding

Neutron transport simulations are ubiquitous in nuclear engineering because they allow one to model experimental systems and render a model platform for easy perturbation of experimental designs. In addition, simulations allow one to gain experimental insight without actually having to go through the trouble of building a physical experiment. Neutron transport simulations can be stochastic or deterministic based. Stochastic neutron transport simulations are typically simulated using the Monte Carlo method and yield very accurate solutions but are computationally expensive, while deterministic methods are typically faster but can be less accurate. Here we focus on optimizing the accuracy of deterministic neutron transport simulations for radiotherapy simulations. Deterministic neutron transport requires discretization of angle, energy, and space to appropriately analyze the system one is trying to model. Discretization of energy is challenging because of the highly variable neutron flux at certain neutron energies. Improper discretization of energy in the transport model can lead to erroneous results and therefore inaccurate interpretations of the solution. In this study, we evaluate Particle Swarm Optimization (PSO) as a mechanism for selecting optimal group structures for radiotherapy shielding. We tested the particle swarm optimization algorithm on radiotherapy shielding problems using Los Alamos National Laboratory's (LANL) main deterministic transport code PARTISN. Results show that the optimized energy group structures generated from the optimization algorithm outperformed LANL's standard energy group structures, and therefore demonstrate utility in using PSO to expedite computation times due to the increased accuracy obtained with a smaller but optimized group structure. (authors)

43 PARTICLE ACCELERATORS↗

Derivative-free optimization methods

In many optimization problems arising from scientific, engineering and artificial intelligence applications, objective and constraint functions are available only as the output of a black-box or simulation oracle that does not provide derivative information. Such settings necessitate the use of methods for derivative-free, or zeroth-order, optimization. We provide a review and perspectives on developments in these methods, with an emphasis on highlighting recent developments and on unifying treatment of such problems in the non-linear optimization and machine learning literature. We categorize methods based on assumed properties of the black-box functions, as well as features of the methods. We first overview the primary setting of deterministic methods applied to unconstrained, non-convex optimization problems where the objective function is defined by a deterministic black-box oracle. We then discuss developments in randomized methods, methods that assume some additional structure about the objective (including convexity, separability and general non-smooth compositions), methods for problems where the output of the black-box oracle is stochastic, and methods for handling different types of constraints.

97 MATHEMATICS AND COMPUTING↗

Model for biological communication in a nanofabricated cell-mimic driven by stochastic resonance

Cells offer natural examples of highly efficient networks of nanomachines. Accordingly, both intracellular and intercellular communication mechanisms in nature are looked to as a source of inspiration and instruction for engineered nanocommunication. Harnessing biological functionality in this manner requires an interdisciplinary approach that integrates systems biology, synthetic biology, and nanofabrication. Recent years have seen the amassing of a tremendous wealth of data from the sequencing of new organisms and from high throughput expression experiments. At the same time, a deeper fundamental understanding of individual cell function has been developed, as exemplified by the growth of fields such as noise biology, which seeks to characterize the role of noise in gene expression. The availability of well characterized biological components coupled with a deeper understanding of cell function has led to efforts to engineer both living cells and to create bio-like functionality in non-living substrates in the field of synthetic biology. Here, we present a model system that exemplifies the synergism between these realms of research. We propose a synthetic gene network for operation in a nanofabricated cell mimic array that propagates a biomolecular signal over long distances using the phenomenon of stochastic resonance. Our system consists of a bacterial quorum sensing signal molecule, a bistable genetic switch triggered by this signal, and an array of nanofabricated cell mimic wells that contain the genetic system. An optimal level of noise in the system helps to propagate a time-varying AHL signal over long distances through the array of mimics. This noise level is determined both by the system volume and by the parameters of the genetic network. Our proposed genetically driven stochastic resonance system serves as a testbed for exploring the potential harnessing of gene expression noise to aid in the transmission of a time-varying molecular signal.

nano communication networks↗