Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “adaptive learning”

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 1,585 records · Page 88

AdaStress

This is a tutorial on AdaStress, a tool for finding and analyzing the likeliest failures in a simulated system under test. The presentation outlines the adaptive stress testing framework, provides a demonstration of use, and showcases several examples of failure detection in a complex real-world system.

Reinforcement learning↗

Dynamic data filtering system and method

A computer-implemented dynamic data filtering system and method for selectively choosing operating data of a monitored asset that modifies or expands a learned scope of an empirical model of normal operation of the monitored asset while simultaneously rejecting operating data of the monitored asset that is indicative of excessive degradation or impending failure of the monitored asset, and utilizing the selectively chosen data for adaptively recalibrating the empirical model to more accurately monitor asset aging changes or operating condition changes of the monitored asset.

97 MATHEMATICS AND COMPUTING↗

Helping Businesses and Communities Better Cope with Extreme Weather

The U.S. Department of Energy's Argonne National Laboratory is helping organizations – including AT&T – and communities better prepare for and adapt to the impacts of climate change and extreme weather events. Argonne's cutting-edge climate science and infrastructure analysis/resilience capabilities have helped a range of organizations (companies, federal agencies and cities) make more informed decisions and minimize negative impacts. To learn how your organization might collaborate with Argonne in this space, contact partners@anl.gov.

54 ENVIRONMENTAL SCIENCES↗

Deep Isolation - Innovation for the Storage and Disposal of High-level Waste and Spent Nuclear Fuel - 19367

Deep Isolation, Inc. is a private, Berkeley based company that has brought innovation to nuclear waste disposal. Deep Isolation has developed a disposal technique that adapts mature drilling technology and customizes it for the purpose of providing a safe and secure solution. Though the stalled nuclear waste program in the US has created unwanted burdens and costs, it has also provided object lessons which Deep Isolation has learned from and designed a novel approach. The solution proposed by Deep Isolation combines the best of the geosciences and social sciences to present the US and other countries with an optimized and tailored approach to disposing of a variety of nuclear waste. This system can complement existing potential storage and disposal sites, or in the case of countries with smaller amounts of waste--the national solution. Deep Isolation innovates where others have stalled and offers a means of deep geological disposal that costs less, takes less time to implement, and cuts down and in some cases eliminates the need for transportation of the waste. The technical solution optimized by Deep Isolation uses established directional drilling technology from the oil and gas industry to drill a vertical drill hole thousands of meters deep and then transition to a horizontal drill hole that is thousands of meters in length. The target media is in or below shale formations, which are present at various depths throughout much of the United States and throughout the world. The availability of suitable geology allows for permanent disposal at multiple locations - including existing nuclear sites, without a need to transport the waste to a central repository. This innovative solution not only offers options for commercial used nuclear fuel and defense spent nuclear fuel and high-level waste, but also other nuclear wastes such as those considered by DoE's past deep bore-hole program, i.e. cesium and strontium capsules. Greater-Than-Class-C (GTCC) waste is also an opportunity for Deep Isolation. DoE's preferred alternative for GTCC includes 'land disposal at generic commercial facilities.' Deep Isolation could provide safe, secure, and permanent disposal at a commercial facility. Deep Isolation complements the technical innovation with best practices from the social sciences. In acknowledgement that the past challenges to implement nuclear waste disposal facilities have been primarily due to socio-political failings, we have designed a means to deeply engage with public and stakeholders. Our approach is committed to partnering with communities, local jurisdictions, states, and Tribes to determine a path forward that benefits all involved, that is fully transparent, collaborative and accountable. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Multiscale graph neural network autoencoders for interpretable scientific machine learning

The goal of this work is to address two limitations in autoencoder-based models: latent space interpretability and compatibility with unstructured meshes. This is accomplished here with the development of a novel graph neural network (GNN) autoencoding architecture with demonstrations on complex fluid flow applications. To address the first goal of interpretability, the GNN autoencoder achieves reduction in the number nodes in the encoding stage through an adaptive graph reduction procedure. Further, this reduction procedure essentially amounts to flowfieldconditioned node sampling and sensor identification, and produces interpretable latent graph representations tailored to the flowfield reconstruction task in the form of so-called masked fields. These masked fields allow the user to (a) visualize where in physical space a given latent graph is active, and (b) interpret the time-evolution of the latent graph connectivity in accordance with the time-evolution of unsteady flow features (e.g. recirculation zones, shear layers) in the domain. To address the goal of unstructured mesh compatibility, the autoencoding architecture utilizes a series of multi-scale message passing (MMP) layers, each of which models information exchange among node neighborhoods at various lengthscales. The MMP layer, which augments standard single-scale message passing with learnable coarsening operations, allows the decoder to more efficiently reconstruct the flowfield from the identified regions in the masked fields. Analysis of latent graphs produced by the autoencoder for various model settings are conducted using unstructured snapshot data sourced from large-eddy simulations in a backward-facing step (BFS) flow configuration with an OpenFOAM-based flow solver at high Reynolds numbers.

97 MATHEMATICS AND COMPUTING↗

Automating Trend Analysis for Spacecraft Constellations

Spacecraft trend analysis is a vital mission operations function performed by satellite controllers and engineers, who perform detailed analyses of engineering telemetry data to diagnose subsystem faults and to detect trends that may potentially lead to degraded subsystem performance or failure in the future. It is this latter function that is of greatest importance, for careful trending can often predict or detect events that may lead to a spacecraft's entry into safe-hold. Early prediction and detection of such events could result in the avoidance of, or rapid return to service from, spacecraft safing, which not only results in reduced recovery costs but also in a higher overall level of service for the satellite system. Contemporary spacecraft trending activities are manually intensive and are primarily performed diagnostically after a fault occurs, rather than proactively to predict its occurrence. They also tend to rely on information systems and software that are oudated when compared to current technologies. When coupled with the fact that flight operations teams often have limited resources, proactive trending opportunities are limited, and detailed trend analysis is often reserved for critical responses to safe holds or other on-orbit events such as maneuvers. While the contemporary trend analysis approach has sufficed for current single-spacecraft operations, it will be unfeasible for NASA's planned and proposed space science constellations. Missions such as the Dynamics, Reconnection and Configuration Observatory (DRACO), for example, are planning to launch as many as 100 'nanospacecraft' to form a homogenous constellation. A simple extrapolation of resources and manpower based on single-spacecraft operations suggests that trending for such a large spacecraft fleet will be unmanageable, unwieldy, and cost-prohibitive. It is therefore imperative that an approach to automating the spacecraft trend analysis function be studied, developed, and applied to missions such as DRACO with the intent that mission operations costs be significantly reduced. The goal of the Constellation Spacecraft Trend Analysis Toolkit (CSTAT) project is to serve as the pathfinder for a fully automated trending system to support spacecraft constellations. The development approach to be taken is evolutionary. In the first year of the project, the intent is to significantly advance the state of the art in current trending systems through improved functionality and increased automation. In the second year, the intent is to add an expert system shell, likely through the adaptation of an existing commercial-off-the-shelf (COTS) or government-off-the-shelf (GOTS) tool to implement some level of the trending intelligence that humans currently provide in manual operations. In the third year, the intent is to infuse the resulting technology into a near-term constellation or formation-flying mission to test it and gain experience in automated trending. The lessons learned from the real missions operations experience will then be used to improve the system, and to ultimately incorporate it into a fully autonomous, closed-loop mission operations system that is truly capable of supporting large constellations. In this paper, the process of automating trend analysis for spacecraft constellations will be addressed. First, the results of a survey on automation in spacecraft mission operations in general, and in trending systems in particular will be presented to provide an overview of the current state of the art. Next, a rule-based model for implementing intelligent spacecraft subsystem trending will be then presented, followed by a survey of existing COTS/GOTS tools that could be adapted for implementing such a model. The baseline design and architecture of the CSTAT system will be presented. Finally, some results obtained from initial software tests and demonstrations will be presented.

Davis, George↗

Development and testing of improved statistical wind power forecasting methods.

Wind power forecasting (WPF) provides important inputs to power system operators and electricity market participants. It is therefore not surprising that WPF has attracted increasing interest within the electric power industry. In this report, we document our research on improving statistical WPF algorithms for point, uncertainty, and ramp forecasting. Below, we provide a brief introduction to the research presented in the following chapters. For a detailed overview of the state-of-the-art in wind power forecasting, we refer to [1]. Our related work on the application of WPF in operational decisions is documented in [2]. Point forecasts of wind power are highly dependent on the training criteria used in the statistical algorithms that are used to convert weather forecasts and observational data to a power forecast. In Chapter 2, we explore the application of information theoretic learning (ITL) as opposed to the classical minimum square error (MSE) criterion for point forecasting. In contrast to the MSE criterion, ITL criteria do not assume a Gaussian distribution of the forecasting errors. We investigate to what extent ITL criteria yield better results. In addition, we analyze time-adaptive training algorithms and how they enable WPF algorithms to cope with non-stationary data and, thus, to adapt to new situations without requiring additional offline training of the model. We test the new point forecasting algorithms on two wind farms located in the U.S. Midwest. Although there have been advancements in deterministic WPF, a single-valued forecast cannot provide information on the dispersion of observations around the predicted value. We argue that it is essential to generate, together with (or as an alternative to) point forecasts, a representation of the wind power uncertainty. Wind power uncertainty representation can take the form of probabilistic forecasts (e.g., probability density function, quantiles), risk indices (e.g., prediction risk index) or scenarios (with spatial and/or temporal dependence). Statistical approaches to uncertainty forecasting basically consist of estimating the uncertainty based on observed forecasting errors. Quantile regression (QR) is currently a commonly used approach in uncertainty forecasting. In Chapter 3, we propose new statistical approaches to the uncertainty estimation problem by employing kernel density forecast (KDF) methods. We use two estimators in both offline and time-adaptive modes, namely, the Nadaraya-Watson (NW) and Quantilecopula (QC) estimators. We conduct detailed tests of the new approaches using QR as a benchmark. One of the major issues in wind power generation are sudden and large changes of wind power output over a short period of time, namely ramping events. In Chapter 4, we perform a comparative study of existing definitions and methodologies for ramp forecasting. We also introduce a new probabilistic method for ramp event detection. The method starts with a stochastic algorithm that generates wind power scenarios, which are passed through a high-pass filter for ramp detection and estimation of the likelihood of ramp events to happen. The report is organized as follows: Chapter 2 presents the results of the application of ITL training criteria to deterministic WPF; Chapter 3 reports the study on probabilistic WPF, including new contributions to wind power uncertainty forecasting; Chapter 4 presents a new method to predict and visualize ramp events, comparing it with state-of-the-art methodologies; Chapter 5 briefly summarizes the main findings and contributions of this report.

17 WIND ENERGY↗

A Barrier-Certificated Reinforcement Learning Approach for Enhancing Power System Transient Stability

Increasing integration of renewable resources brings more flexibility and poses new challenges to modern power systems, leading to highly nonlinear and complex dynamics. Here, this paper aims to provide a general solution framework to traditional control problems, such as frequency control and voltage control, which attempt to maintain the stability of either synchronous generators-governed or inverter-governed systems when subjected to a disturbance and simultaneously guarantee operational constraints, providing a complete complement to existing works on control design. Building on reinforcement learning (RL) and control barrier functions, the framework includes two subsystems, i.e., a model-free controller and a barrier-certification system, which discover RL-based control actions and sequentially filter them using a barrier certificate to satisfy operational constraints. Calculating a barrier function is generally challenging for a complex power system. This is addressed by representing the barrier function using neural networks (NNs) and data-based approaches. An adaptive method is introduced to certify the neural barrier function that perseveres barrier conditions, which is more compatible with online implementation. The proposed framework synthesizes a stabilizing controller that satisfies predefined safety regions. The effectiveness of the proposed framework is demonstrated via several comparative case studies.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Machine Learning Toolkit for Extreme Scale

Support Vector Machines (SVM) is a popular machine learning technique, which has been applied to a wide range of domains such as science, finance, and social networks for supervised learning. MaTEx undertakes the challenge of designing a scalable parallel SVM training algorithm for large scale systems, which includes commodity multi-core machines, tightly connected supercomputers and cloud computing systems. Several techniques are proposed for improved speed and memory space usage including adaptive and aggressive elimination of samples for faster convergence , and sparse format representation of data samples. Several heuristics for earliest possible to lazy elimination of non-contributing samples are considered in MaTEx. In many cases, where an early sample elimination might result in a false positive, low overhead mechanisms for reconstruction of key data structures are proposed. The proposed algorithm and heuristics are implemented and evaluated on various publicly available datasets

Source record↗

The application of model-referenced adaptive control to robotic manipulators

The achievement of quality dynamic performance in manipulator systems is difficult using conventional control methods because of both the inherent geometric nonlinearities of these systems and the dependence of the system dynamics on the characteristics of manipulated objects. A model-referenced adaptive control law is developed for maintaining uniformly good performance over a wide range of motions and payloads. The effectiveness of the approach is demonstrated in several simulations and the system stability as a function of input is investigated. Also developed is a 'learning signal' approach designed to minimize initial transients arising from abrupt changes in the inertial payload.

Dubowsky, S.↗

Supporting Development of Satellite's Guidance Navigation and Control Software: A Product Line Approach

The NASA Goddard Space Flight Center Flight Software Branch (FSB) is developing a Guidance, Navigation, and Control (GNC) Flight Software (FSW) product line. The demand for increasingly more complex flight software in less time while maintaining the same level of quality has motivated us to look for better FSW development strategies. The GNC FSW product line has been planned to address the core GNC FSW functionality very similar on many recent low/near Earth missions in the last ten years. Unfortunately these missions have not accomplished significant drops in development cost since a systematic approach towards reuse has not been adopted. In addition, new demands are continually being placed upon the FSW which means the FSB must become more adept at providing GNC FSW functionality's core so it can accommodate additional requirements. These domain features together with engineering concepts are influencing the specification, description and evaluation of FSW product line. Domain engineering is the foundation for emerging product line software development approaches. A product line is 'A family of products designed to take advantage of their common aspects and predicted variabilities'. In our product line approach, domain engineering includes the engineering activities needed to produce reusable artifacts for a domain. Application engineering refers to developing an application in the domain starting from reusable artifacts. The focus of this paper is regarding the software process, lessons learned and on how the GNC FSW product line manages variability. Existing domain engineering approaches do not enforce any specific notation for domain analysis or commonality and variability analysis. Usually, natural language text is the preferred tool. The advantage is the flexibility and adapt ability of natural language. However, one has to be ready to accept also its well-known drawbacks, such as ambiguity, inconsistency, and contradictions. While most domain analysis approaches are functionally oriented, the idea of applying the object-oriented approach in domain analysis is not new. Some authors propose to use UML as the notation underlying domain analysis. Our work is based on the same idea of merging UML and domain analysis. Further, we propose a few extensions to UML in order to express variability, and we define precisely their semantics so that a tool can support them. The extensions are designed to be implemented on the API of a popular industrial CASE tool, with obvious advantages in cost and availability of tool support. The paper outlines the product line processes and identifies where variability must be addressed. Then it describes the product line products with respect to how they accommodate variability. The Celestial Body subdomain is used as a working example. Our results to date are summarized and plans for the future are described.

McComas, David↗

Intelligent Control for the BEES Flyer

This paper describes the effort to provide a preliminary capability analysis and a neural network based adaptive flight control system for the JPL-led BEES aircraft project. The BEES flyer was envisioned to be a small, autonomous platform with sensing and control systems mimicking those of biological systems for the purpose of scientific exploration on the surface of Mars. The platform is physically tightly constrained by the necessity of efficient packing within rockets for the trip to Mars. Given the physical constraints, the system is not an ideal configuration for aerodynamics or stability and control. The objectives of this effort are to evaluate the aerodynamics characteristics of the existing design, to make recommendaaons as to potential improvements and to provide a control system that stabilizes the existing aircraft for nominal flight and damaged conditions. Towards this several questions are raised and analyses are presented to arrive at answers to some of the questions raised. CART3D, a high-fidelity inviscid analysis package for conceptual and preliminary aerodynamic design, was used to compute a parametric set of solutions over the expected flight domain. Stability and control derivatives were extracted from the database and integrated with the neural flight control system. The Integrated Vehicle Modeling Environment (IVME) was also used for estimating aircraft geometric, inertial, and aerodynamic characteristics. A generic neural flight control system is used to provide adaptive control without the requirement for extensive gain scheduling or explicit system identification. The neural flight control system uses reference models to specify desired handling qualities in the roll, pitch, and yaw axes, and incorporates both pre-trained and on-line learning neural networks in the inverse model portion of the controller. Results are presented for the BEES aircraft in the subsonic regime for terrestrial and Martian environments.

Krishnakumar, K.↗

NASA's Morphing Project Research Summaries in Fiscal Year 2002

The Morphing Project at the National Aeronautics and Space Agency s (NASA) Langley Research Center (LaRC) is part of the Breakthrough Vehicle Technologies Project, Vehicle Systems Program that conducts fundamental research on advanced technologies for future flight vehicles. The objectives of the Morphing Project are to develop and assess the advanced technologies and integrated component concepts to enable efficient, multi-point adaptability of flight vehicles; primarily through the application of adaptive structures and adaptive flow control to substantially alter vehicle performance characteristics. This document is a compilation of research summaries and other information on the project for fiscal year 2002. The focus is to provide a brief overview of the project content, technical results and lessons learned. At the time of publication, the Vehicle Systems Program (which includes the Morphing Project) is undergoing a program re-planning and reorganization. Accordingly, the programmatic descriptions of this document pertain only to the program as of fiscal year 2002.

McGowan, Anna-Maria R.↗

Climate Change Impacts and Responses: Societal Indicators for the National Climate Assessment

The Climate Change Impacts and Responses: Societal Indicators for the National Climate Assessment workshop, sponsored by the National Aeronautics and Space Administration (NASA) for the National Climate Assessment (NCA), was held on April 28-29, 2011 at The Madison Hotel in Washington, DC. A group of 56 experts (see list in Appendix B) convened to share their experiences. Participants brought to bear a wide range of disciplinary expertise in the social and natural sciences, sector experience, and knowledge about developing and implementing indicators for a range of purposes. Participants included representatives from federal and state government, non-governmental organizations, tribes, universities, and communities. The purpose of the workshop was to assist the NCA in developing a strategic framework for climate-related physical, ecological, and socioeconomic indicators that can be easily communicated with the U.S. population and that will support monitoring, assessment, prediction, evaluation, and decision-making. The NCA indicators are envisioned as a relatively small number of policy-relevant integrated indicators designed to provide a consistent, objective, and transparent overview of major variations in climate impacts, vulnerabilities, adaptation, and mitigation activities across sectors, regions, and timeframes. The workshop participants were asked to provide input on a number of topics, including: (1) categories of societal indicators for the NCA; (2) alternative approaches to constructing indicators and the better approaches for NCA to consider; (3) specific requirements and criteria for implementing the indicators; and (4) sources of data for and creators of such indicators. Socioeconomic indicators could include demographic, cultural, behavioral, economic, public health, and policy components relevant to impacts, vulnerabilities, and adaptation to climate change as well as both proactive and reactive responses to climate change. Participants provided inputs through in-depth discussion in breakout sessions, plenary sessions on break-out results, and several panels that provided key insights about indicators, lessons learned through experience with developing and implementing indicators, and thoughts on how the NCA could proceed to develop indicators for the NCA.

Kenney, Melissa A.↗

Spaceborne Lidar Retrievals of PM2.5 for Air Quality Studies and Applications

Fine particulate matter (PM2.5) substantially contributes to air pollution and negatively affects human health. While many studies have investigated the use of passive column-integrated aerosol optical depth to infer surface PM2.5, the use of lidar observations for air quality characterization is not nearly as extensive. Lidar measurements are critical, however, due to the vertical aerosol information they provide, including near the surface. In this presentation, we first provide an overview of various lidar-based approaches for estimating PM2.5 concentrations and then discuss how lidar measurements can assist other air quality applications. For example, estimates of PM2.5 have been obtained in a physics-based approach through CALIOP near-surface aerosol extinction retrievals, assumptions on the mass extinction efficiency, and incorporating other parameters (an aerosol hygroscopic growth factor and PM2.5/PM10 ratio). Application of this algorithm over the contiguous United States (CONUS) from 2006 to 2018 yielded larger PM2.5 values over the eastern and western CONUS (~10-15 μg/m³) and lower PM2.5 levels in the central CONUS (~5 μg/m³). These spatial patterns were similar to those from gridded PM2.5 concentrations obtained through in situ measurements at ground stations operated by the US Environmental Protection Agency. In another approach, the Cloud Aerosol Transport System (CATS) lidar was used with the Goddard Earth Observing System (GEOS) model in a 1D ensemble-based variational technique to obtain PM2.5 over the US and Europe, and the spatial patterns of the CATS/GEOS based PM2.5 concentrations generally captured those from surface stations (with corresponding hourly EPA PM2.5 vs CATS PM2.5 statistics of R=0.4 and bias=1.5 μg/m³). In our recent work, as part of the Models, In situ, and Remote sensing of Aerosols (MIRA) Working Group, we have applied both the CALIOP and CATS/GEOS based approaches over the highly polluted country of India during the post-monsoon season (September-October 2016). We derived elevated levels of two-month mean PM2.5 (~100 μg/m³) in northern India, especially near New Delhi. These high PM2.5 concentrations in the Indo-Gangetic plain are driven in large part from the seasonal burning of crop residue and meteorological conditions typical at this time of the year, such as low wind speeds and a shallow boundary layer. While the satellite-derived PM2.5 moderately replicates (R = ~0.7-0.9) the spatial variability in the two-month mean of surface in situ PM2.5 from monitoring sites operated by the Central and State Pollution Control Boards, we show results from specific scenes for which there are large deviations between the satellite-derived PM2.5 and in situ measurements. Other current work on this topic focuses on developing PM2.5 estimates using airborne high spectral resolution lidar measurements through machine learning regression algorithms and involves several parameters (e.g., aerosol extinction, color ratio, lidar ratio). Application of this method over major metropolitan areas in the US and Asia have resulted in high correlations (R = 0.93) with surface measurements. This airborne lidar approach can be adapted to spaceborne lidar measurements, and all three of these approaches can be applied to ESA’s EarthCARE Atmospheric Lidar instrument, setting the stage for the future Cloud Aerosol Lidar for Global Scale Observations of the Ocean-Land Atmosphere System (CALIGOLA) mission. Ultimately, beyond estimates of PM2.5, the aerosol vertical distribution from lidars can benefit studies involving passive sensor approaches for PM2.5 proxies (including from geostationary satellites), wildfire smoke plume injection heights, volcanic emissions (e.g., ash height retrievals), and aerosol/air quality model assimilation, evaluation, and forecasts.

Travis D Toth↗

Machine Learning Toolkit for Extreme Scale

Support Vector Machines (SVM) is a popular machine learning technique, which has been applied to a wide range of domains such as science, finance, and social networks for supervised learning. MaTEx undertakes the challenge of designing a scalable parallel SVM training algorithm for large scale systems, which includes commodity multi-core machines, tightly connected supercomputers and cloud computing systems. Several techniques are proposed for improved speed and memory space usage including adaptive and aggressive elimination of samples for faster convergence , and sparse format representation of data samples. Several heuristics for earliest possible to lazy elimination of non-contributing samples are considered in MaTEx. In many cases, where an early sample elimination might result in a false positive, low overhead mechanisms for reconstruction of key data structures are proposed. The proposed algorithm and heuristics are implemented and evaluated on various publicly available datasets

Source record↗

Meta-modeling strategy for data-driven forecasting

Accurately forecasting the weather is a key requirement for climate change mitigation. Data-driven methods offer the ability to make more accurate forecasts, but lack interpretability and can be expensive to train and deploy if models are not carefully developed. Here, we make use of two historical climate data sets and tools from machine learning, to accurately predict temperature fields. Furthermore, we are able to use low fidelity models that are cheap to train and evaluate, to selectively avoid expensive high fidelity function evaluations, as well as uncover seasonal variations in predictive power. This allows for an adaptive training strategy for computationally efficient geophysical emulation.

Skinner, Dominic J↗

Reinforcement Learning for Load-balanced Parallel Particle Tracing

We explore an online reinforcement learning (RL) paradigm to dynamically optimize parallel particle tracing performance in distributed-memory systems. Our method combines three novel components: (1) a work donation algorithm, (2) a high-order workload estimation model, and (3) a communication cost model. First, we design an RL-based work donation algorithm. Our algorithm monitors workloads of processes and creates RL agents to donate data blocks and particles from high-workload processes to low-workload processes to minimize program execution time. The agents learn the donation strategy on the fly based on reward and cost functions designed to consider processes' workload changes and data transfer costs of donation actions. Second, we propose a workload estimation model, helping RL agents estimate the workload distribution of processes in future computations. Third, we design a communication cost model that considers both block and particle data exchange costs, helping RL agents make effective decisions with minimized communication costs. We demonstrate that our algorithm adapts to different flow behaviors in large-scale fluid dynamics, ocean, and weather simulation data. Our algorithm improves parallel particle tracing performance in terms of parallel efficiency, load balance, and costs of I/O and communication for evaluations with up to 16,384 processors.

Distributed and parallel particle tracing↗