Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Agent based modeling”

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 109 records · Page 6
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

Evaluating the impacts of Variable Message Signs on Airport Curbside Performance Using Microsimulation

Curbs play a vital role in facilitating vehicle access and egress for individuals at airports. Inefficiently allocating this resource hinders airport accessibility and productivity, resulting in congestion, longer travel times, and increased pollution. As airport demand fluctuates throughout the day and grows over time, airports face intensified curbside pressure. Yet, curb management research is significantly less robust at airports than in urban areas. Given the unbalanced nature of airport demand—riders tend to arrive simultaneously at specific entrances at certain hours—Variable Message Sign (VMS) arises as a cost-effective technology to divert vehicles from congested to underutilized curbs. Still, VMS implementation faces a significant challenge. Historically, airports have managed VMS heuristically and by intuition rather than an evidence-based approach. This research investigates the impacts of implementing VMS on curb performance at airports. By considering different driver compliance rates (DCR), we aim to determine when the sign should be turned on and off to diverge traffic to avoid undesired externalities while enhancing curb performance. Using a validated agent-based microsimulation model, VISSIM, we analyzed the Seattle-Tacoma (SeaTac) Airport as a case study. We modeled sixteen VMS management scenarios and a baseline where the message sign is not displayed, diverging vehicles between the departures and arrivals access levels at four different moments (early morning, morning, afternoon, and late night). We quantified the effects of VMS using seven metrics, including curb productivity index (CPI), curb accessibility (CA), queue length, queue duration, delay, vehicle counts, and emissions. The results of each scenario were compared against the baseline using absolute and relative changes and Repeated Measures ANOVA. Overall, VMS improved curb performance and traffic conditions at the airport, reducing emissions by 14.8% to 8.9%. Moreover, significant reductions in queue length (1,150 ft to 100 ft) and duration (15 to 144 minutes) were observed in the sending link under all VMS policies. However, impacts on the receiving link varied based on congestion, with significant increases in queue duration (9.8 to 24 min) when congested but no substantial changes in free flow. Notably, diverging vehicles to congested links resulted in non-significant results, and activating late and deactivating late VMS affected curb productivity (-5.8% to -61.4%), curb accessibility (-16.5% to -25.8%), cumulative counts (-33.4% to -59.4%), and vehicle delay (95.98% to 594.3%). Activating VMS before congestion begins in the sending link and deactivating before a queue forms in the receiving link yield the most significant improvements: 8.1% to 10.1% in CPI, 9.4% to9.6% in CA, -29.3% to -77.9% in total delay, -11.6% to -13.9% in total emissions, and 101% to 103% in cumulative counts. As the analysis was made with a wide range of time periods, access levels, driver compliance rates, and scenarios, we believe our findings can provide valuable insights into how airports should manage VMS. Our work introduces a novel approach to the scientific airport literature, as some of our metrics were previously unexplored. Additionally, we propose a methodology that other airports can adopt to maximize their curb performance.

Gutierrez, Jorge D.↗
DOE OSTI2025

Data Assimilation for Robust UQ Within Agent-Based Simulation on HPC Systems

Agent-based simulation provides a powerful tool for in silico system modeling. However, these simulations do not provide built-in methods for uncertainty quantification (UQ). Within these types of models a typical approach to UQ is to run multiple realizations of the model then compute aggregate statistics. This approach is limited due to the compute time required for a solution. When faced with an emerging biothreat, public health decisions need to be made quickly and solutions for integrating near real-time data with analytic tools are needed. We propose an integrated Bayesian UQ framework for agent-based models based on sequential Monte Carlo sampling. Given streaming or static data about the evolution of an emerging pathogen this Bayesian framework provides a distribution over the parameters governing the spread of a disease through a population. These estimates of the spread of a disease may be provided to public health agencies seeking to abate the spread. By coupling agent-based simulations with Bayesian modeling in a data assimilation, our proposed framework provides a powerful tool for modeling dynamical systems in silico. We propose a method which reduces model error and provides a range of realistic possible outcomes. Moreover, our method addresses two primary limitations of ABMs: the lack of UQ and an inability to assimilate data. Our proposed framework combines the flexibility of an agent-based model with UQ provided by the Bayesian paradigm in a workflow which scales well to HPC systems. We provide algorithmic details and results on a simulated outbreak with both static and streaming data.

Spannaus, Adam [ORNL] (ORCID:0000000225213657)↗
DOE OSTI2024

Modeling Electric Vehicle Charging Load Using Origin-Destination Data

The accelerating adoption of electric vehicles (EVs) poses challenges to the power grid, necessitating precise representation of mobility patterns for effective infrastructure upgrades. Traditional simulation-based charging demand estimation faces limitations in generating trip chains reflective of actual travel patterns without complex network modeling. Hence, an innovative agent-based trip chain generation model is introduced to overcome these challenges. Drawing from the National Household Travel Survey (NHTS) and the NextGen NHTS origin-destination add-on data for Clarke County, Georgia, this study proposes a simulation method capturing both temporal and spatial mobility patterns without relying on extensive network topology data. The resulting trip chains predict EV charging load at the Census Block Group level, validated with a 1.03 correlation to actual trip counts, affirming their reflective accuracy. Two charging scenarios, residential-only and charging-everywhere, reveal distinct demand profiles. The charging-everywhere scenario aligns closely with the trip profile, while the residential-only scenario exhibits an afternoon peak slightly surpassing the former. This study contributes a data-driven charging demand estimation methodology, offering critical insights for grid resiliency planning amid the evolving landscape of EV adoption.

Pan, Melrose↗
NASA NTRS1999

Remote Agent Demonstration

We describe the computer demonstration of the Remote Agent Experiment (RAX). The Remote Agent is a high-level, model-based, autonomous control agent being validated on the NASA Deep Space 1 spacecraft.

Dorais, Gregory A.↗
NASA NTRS1990

A knowledge base architecture for distributed knowledge agents

A tuple space based object oriented model for knowledge base representation and interpretation is presented. An architecture for managing distributed knowledge agents is then implemented within the model. The general model is based upon a database implementation of a tuple space. Objects are then defined as an additional layer upon the database. The tuple space may or may not be distributed depending upon the database implementation. A language for representing knowledge and inference strategy is defined whose implementation takes advantage of the tuple space. The general model may then be instantiated in many different forms, each of which may be a distinct knowledge agent. Knowledge agents may communicate using tuple space mechanisms as in the LINDA model as well as using more well known message passing mechanisms. An implementation of the model is presented describing strategies used to keep inference tractable without giving up expressivity. An example applied to a power management and distribution network for Space Station Freedom is given.

Riedesel, Joel↗
NASA NTRSDate not supplied

SafeMAP: Safe Multi-Agent Planning Framework Based on Dynamic Probabilistic Risk Assessment

This paper proposes a risk-aware framework for Safe Multi-Agent Planning (SafeMAP) that unifies disparate models for multi-agent systems in a Markovian process that allows for simultaneous system health monitoring, decision making under uncertainty, and multi-agent system collaboration. As operations beyond low earth orbit mature, there is an increased need for autonomous cyber-physical systems with onboard decision making capabilities. Multi-agent cyber-physical systems in particular offer the potential of increased efficiency, resiliency, and mission capabilities for future applications such as multi-rover terrain operations, distributed satellite operations, and management of smart lunar habitats. SafeMAP utilizes physics-based models of each agent and the relevant components, probability models of the environment and component operational states, and reward models for mission-specific objectives such as scientific task completion or resource consumption. The output of SafeMAP is a set of mission plans that satisfy the mission objective under specified risk/reward constraints. A readable interpretation of each of these generated mission plans is provided as an additional output. SafeMAP has been demonstrated on a simulated case study involving a four-rover system performing surface mapping operations and science tasks. Results of this paper demonstrate SafeMAP’s ability to generate explainable mission plans that satisfy the mission objective while minimizing risk under nominal and off-nominal conditions.

Mohammad Hejase↗
NASA NTRSDate not supplied

SafeMAP: Safe Multi-Agent Planning framework based on Dynamic Probabilistic Risk Assessment

This paper proposes a risk-aware framework for Safe Multi-Agent Planning (SafeMAP) that unifies disparate models for multi-agent systems in a Markovian process that allows for simultaneous system health monitoring, decision making under uncertainty, and multi-agent system collaboration. As operations beyond low earth orbit mature, there is an increased need for autonomous cyber-physical systems with onboard decision making capabilities. Multi-agent cyber-physical systems in particular offer the potential of increased efficiency, resiliency, and mission capabilities for future applications such as multi-rover terrain operations, distributed satellite operations, and management of smart lunar habitats. SafeMAP utilizes physics-based models of each agent and the relevant components, probability models of the environment and component operational states, and reward models for mission-specific objectives such as scientific task completion or resource consumption. The output of SafeMAP is a set of mission plans that satisfy the mission objective under specified risk/reward constraints. A readable interpretation of each of these generated mission plans is provided as an additional output. SafeMAP has been demonstrated on a simulated case study involving a four-rover system performing surface mapping operations and science tasks. Results of this paper demonstrate SafeMAP’s ability to generate explainable mission plans that satisfy the mission objective while minimizing risk under nominal and off-nominal conditions.

Mohammad Hejase↗
DOE OSTI2024

Machine learning for the identification of phase transitions in interacting agent-based systems: A Desai-Zwanzig example

Deriving closed-form analytical expressions for reduced-order models, and judiciously choosing the closures leading to them, has long been the strategy of choice for studying phase- and noise-induced transitions for agent-based models (ABMs). In this paper, we propose a data-driven framework that pinpoints phase transitions for an ABM—the Desai-Zwanzig model—in its mean-field limit, using a smaller number of variables than traditional closed-form models. To this end, we use the manifold learning algorithm Diffusion Maps to identify a parsimonious set of data-driven latent variables, and we show that they are in one-to-one correspondence with the expected theoretical order parameter of the ABM. We then utilize a deep learning framework to obtain a conformal reparametrization of the data-driven coordinates that facilitates, in our example, the identification of a single parameter-dependent ordinary differential equation (ODE) in these coordinates. Additionally, we identify this ODE through a residual neural network inspired by a numerical integration scheme (forward Euler). We then use the identified ODE—enabled through an odd symmetry transformation—to construct the bifurcation diagram exhibiting the phase transition.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Model-based Hierarchical Reinforcement Learning for Improved Physical Security Design: A Prototype

Prior work in FY24 developed an adversarial AI agent aid in path analysis of physical protection systems. This agent, trained using a model-based reinforcement learning algorithm, was able to successfully learn the most vulnerable path in facilities. It was able to extend the current state of practice for physical protection design by exhibiting dynamic behavior based on current environmental conditions. Whereas PathTrace largely performs a static, graph-based analysis, the AI agent was able to make decisions based on relative position in the facility, current conditions (was the adversarial agnet discovered?), and proximity to secondary targets. The agent demonstrated some novel capabilities, but had limitations that need to be resolved before it can be used for production purposes. For example, the adversarial agent generalizes poorly and takes a relatively long time to train. Nonetheless, there is still considerable promise for developing the adversarial agent further in order to explore even richer, more dynamic behaviors (e.g., adversary motivations, environmental debris, and more). This work considers a complementary idea; development of a planning agent. The planning agent is envisioned as an auto-complete-like tool that can help accelerate security system design by human experts. The agent would respect existing barriers and sensors placed by a human expert while offering cost-effective suggestions (i.e., implicitly balancing effectiveness with cost) to improve the design. The goal is for this agent to be part of an expert’s toolbox, not to totally upend the current state-of-practice, or to displace human experts. The ultimate goal would be concurrent training of both the adversarial and planning agent together, to learn entirely through self-play. This would represent an entirely new way of performing system deign. We selected a hierarchical, model-based reinforcement learning algorithm to serve as the planning agent. This is an extension of concepts used in the prior FY24 adversarial agent work. There, we had a single agent acting an environment. Here, we have two different sub-agents (policies), working together, to form a complete agent. There is a manager policy, which can select abstract goals on slower time scales, and a worker, which performs primitive actions to reach goals selected by the manager. It is worth noting that this class of algorithm is challenging to work with. From our understanding, our work is one of the first successful uses of model-based reinforcement learning (MBRL) in nuclear energy1 , and likely the first hierarchical model-based reinforcement learning application in nuclear energy. Further, this work is one of the first known attempts to apply AI to perform a design tasks in nuclear energy. Consequently, there were significant implementation challenges and the bulk of the work was focused on successful implementation and algorithm design. The results presented here are very low technology readiness level as a consequence of the lack of related literature, but still represent a significant step forward in the pursuit of applied AI for design.

42 ENGINEERING↗
NASA NTRS2020

Multi-Agent Motion Planning using Deep Learning for Space Applications

State-of-the-art motion planners cannot scale to a large number of systems. Motion planning for multiple agents is an NP (non-deterministic polynomial-time) hard problem, so the computation time increases exponentially with each addition of agents. This computational demand is a major stumbling block to the motion planner's application to future NASA missions involving the swarm of space vehicles. We applied a deep neural network to transform computationally demanding mathematical motion planning problems into deep learning-based numerical problems. We showed optimal motion trajectories can be accurately replicated using deep learning-based numerical models in several 2D and 3D systems with multiple agents. The deep learning-based numerical model demonstrates superior computational efficiency with plans generated 1000 times faster than the mathematical model counterpart.

Madani, Ramtin↗
DOE OSTI2025

Propagating synthetic populations with dynamic Bayesian networks: a framework for long-horizon demographic forecasting

This study presents a dynamic demographic microsimulator using dynamic Bayesian networks to forecast long–term changes in household and individual life events. Leveraging longitudinal Panel Study of Income Dynamics (PSID) data, two networks for individuals and households were modeled to simulate transitions in employment, income, education, marriage, childbirth, leaving the parental home, home ownership, mortality, and household formation or dissolution. Across 1,000 simulation runs spanning 24 years, household–level outcomes remain highly accurate and individual–level predictions reasonable. Although accuracy naturally declines with projection horizon, performance remains promising at both levels. This study addresses a key limitation of existing population synthesis models, which typically generate only a single static snapshot of the population. In conclusion, by introducing a framework that propagates cross-sectional outputs into the future, the microsimulator enables the tracking of demographic evolution over time, enhances realism in population-based simulations, and supplies credible inputs to agent-based travel demand models.

Demographic modeling↗
DOE OSTI2024

A Distributed Model Identification Algorithm for Multi-Agent Systems: Preprint

In this study, we investigate agent-based approach for system model identification with emphasis on power distribution system applications. Departing from conventional practices of relying on historical data for offline model identification, we adopt online update approach utilizing real-time data by employing the latest data points for gradient computation. This methodology offers advantages including a large reduction in the communication network's bandwidth requirements by minimizing the data exchanged at each iteration and enabling the model to adapt in real-time to disturbances. Furthermore, we extend our model identification process from linear frameworks to more complex non-linear convex models. This extension is validated through numerical studies demonstrating improved control performance for a synthetic IEEE test case.

data-driven control↗
DOE OSTI2026

Study of wrap mode impact on Pseudomonas aeruginosa motion in the chemotactic field of fungi

An agent-based discrete computational model biologically calibrated to Pseudomonas aeruginosa migration is used to explore the impacts of bacterial reversals and wrap mode on the efficiency of motion in different environments, both with and without chemotaxis. It is first shown that wrap mode increases the exploration of continuous multimodal chemotactic profiles such as those produced by biologically relevant fungal networks. For cells undergoing a run-reverse pattern, it is shown that the bacteria are likely to remain at the first local chemoattractant maximal production site on a hypha they find. However, with wrap mode, the bacteria can more easily escape these local sites to further explore their neighboring environment along the fungi, suggesting that wrap mode may be beneficial for migration along the fungi in liquid. In a different set of simulations of bacterial motion close to an isolated chemotactic source, wrap mode is shown to increase the ability of a bacterium to reorient toward the source while reducing the overall motion required for similar chemotactic efficiency as a run-reverse strategy, suggesting a potential metabolic benefit. In contrast, model simulations show that wrap mode can increase the rate of dispersal of P. aeruginosa in a nonchemotactic environment.

Hansen, Austin [University of California, Riversid↗
DOE OSTI2025

Economic Storage Size Optimization for Electric Vehicle Extreme-Fast Charging Stations

En-route charging infrastructure for electric vehicles is critical to support transportation needs. These charging stations are likely to have high loads and especially sharp peak loads given fast charging capabilities needed to meet transportation schedules. In order to reduce both strain on distribution grid infrastructure and charging station operational costs, many stations are likely to employ behind the meter storage. This paper demonstrates a behind the meter storage sizing optimization that employs an open-source agent-based vehicle behavior model (BEAM) to determine the best sizing across many scenarios. This optimization and analysis is novel in that it examines how storage size impacts not only charging station cost and peak load, but also vehicle queue times. The optimization is also applied across a wide analysis region with sufficient diversity and numbers to provide novel statistical analysis of optimal sizes.

Aka, Julius↗
DOE OSTI2024

Firm Synthesizer and Supply-chain Simulator (SynthFirm) v1.0

SynthFirm is a large-scale agent-based freight demand model which generates a complete synthetic population of firms in the U.S. and the business-to-business commodity flows between them. Using publicly available data sources as inputs, SynthFirm simulates detailed firm and fleet characteristics, commodity production and consumption, formation of supply chains, and selection of shipping modes, all of which are essential drivers of commodity flow at a disaggregate level.

Xu, Xiaodan↗
DOE OSTI2025

Firm Synthesizer and Supply-chain Simulator (SynthFirm) v2.0

SynthFirm is a national-scale agent-based freight demand model which generates a complete synthetic population of firms in the U.S. and the business-to-business commodity flows between them. Using publicly available data sources as inputs, SynthFirm simulates detailed firm and fleet characteristics, commodity production and consumption, formation of supply chains, and selection of shipping modes, all of which are essential drivers of commodity flow at a disaggregate level. The SynthFirm 2.0 version includes national commercial vehicle fleet generation, international trade simulation and automized model validation pipeline, which allows seemless deployment across the nation and build a comprehensive freight inventories at national scale or for selected region.

Yang, Hung-Chia [Lawrence Berkeley National Labora↗
NASA NTRS2003

An Agent Inspired Reconfigurable Computing Implementation of a Genetic Algorithm

Many software systems have been successfully implemented using an agent paradigm which employs a number of independent entities that communicate with one another to achieve a common goal. The distributed nature of such a paradigm makes it an excellent candidate for use in high speed reconfigurable computing hardware environments such as those present in modem FPGA's. In this paper, a distributed genetic algorithm that can be applied to the agent based reconfigurable hardware model is introduced. The effectiveness of this new algorithm is evaluated by comparing the quality of the solutions found by the new algorithm with those found by traditional genetic algorithms. The performance of a reconfigurable hardware implementation of the new algorithm on an FPGA is compared to traditional single processor implementations.

Weir, John M.↗
DOE OSTI2026

The ballad of LLM agents: philosophical reasoning for chemistry

Large language models (LLMs) show remarkable potential for scientific reasoning but often produce unreliable or scientifically unactionable outputs when faced with multi-step logic, domain grounding, and interpretability challenges, especially in complex fields like chemistry and materials science. Here, we introduce a framework of philosophical reasoning agents, inspired by canonical thinkers such as Socrates, Descartes, Kant, and Hume, to guide LLM behavior via structured prompt engineering. These agents embody distinct reasoning paradigms (dialectical inquiry, deductive logic, rule-based judgment, and empirical validation) and are evaluated across multiple chemistry subdomains, physical, analytical, general, inorganic, and organic chemistry, using the ChemBench benchmark. Our agentic prompting approach yields substantial accuracy gains on open-ended numerical chemistry questions, with gains of +11.5 percentage points for GPT-4o with Hume, +4.5 percentage points for GPT-5 with Kant, and +21.8 percentage points for GPT-5.1 with Socrates at the strict 1% error threshold, relative to the corresponding base models. Beyond accuracy, we observe benchmark-level model–agent performance patterns, suggesting that different prompting styles interact differently with each base model. These findings demonstrate that embedding philosophy-of-science principles into multi-agent frameworks can improve and produce interpretable, adaptive, and domain-aligned scientific LLMs.

Harb, Hassan [Argonne National Laboratory (ANL), A↗