Results for “Agent based modeling” — Page 15

DOE OSTI2026

From Rules to Reasoning: A Survey of Large Language Model-Based Approaches to Scientific Hypothesis and Idea Generation

Scientific hypothesis generation represents a fundamental challenge in contemporary research due to exponentially expanding literature volumes and increasing disciplinary specialization. Large language models (LLMs) have emerged as transformative tools for automated scientific discovery, moving beyond traditional rule-based and literature-mining approaches. Four paradigmatic approaches define current LLM-driven hypothesis generation: direct prompting and fine-tuning methods, knowledge-enhanced frameworks integrating retrieval-augmented generation (RAG), multi-agent collaborative systems simulating research teams, and reasoning-focused approaches implementing cognitive architectures. Domain-specific applications demonstrate statistical equivalence to human expert performance in social psychology, experimental validation in biomedical research, and near-expert quality in astronomy. Evaluation methodologies encompass human expert assessment, LLM-as-judge frameworks, and comprehensive benchmarking systems. Technical challenges include hallucination management, knowledge integration limitations, and balancing novelty with feasibility. Future directions emphasize hybrid neural-symbolic architectures and sophisticated human-AI collaboration models for responsible scientific discovery acceleration.

AI-driven discovery↗

NASA NTRS1985

Extension of Ostwald Ripening Theory

The objective is to develop models based on the mean field approximation of Ostwald ripening to describe the growth of second phase droplets or crystallites. The models will include time variations in nucleation rate, control of saturation through addition of solute, precipitating agents, changes in temperature, and various surface kinetic effects. Numerical integration schemes have been developed and tested against the asymptotic solution of Liftshitz, Slyozov and Wagner (LSW). A second attractor (in addition to the LSW distribution) has been found and, contrary to the LSW theory, the final distribution is dependent on the initial distribution. A series of microgravity experiments is being planned to test this and other results from this work.

Baird, J.↗

NASA NTRS1995

Use of agents to implement an integrated computing environment

Integrated Product and Process Development (IPPD) embodies the simultaneous application to both system and quality engineering methods throughout an iterative design process. The use of IPPD results in the time-conscious, cost-saving development of engineering systems. To implement IPPD, a Decision-Based Design perspective is encapsulated in an approach that focuses on the role of the human designer in product development. The approach has two parts and is outlined in this paper. First, an architecture, called DREAMS, is being developed that facilitates design from a decision-based perspective. Second, a supporting computing infrastructure, called IMAGE, is being designed. Agents are used to implement the overall infrastructure on the computer. Successful agent utilization requires that they be made of three components: the resource, the model, and the wrap. Current work is focused on the development of generalized agent schemes and associated demonstration projects. When in place, the technology independent computing infrastructure will aid the designer in systematically generating knowledge used to facilitate decision-making.

Hale, Mark A.↗

NASA NTRS1999

The Extinction of Low Strain Rate Diffusion Flames by a Suppressant

This paper describes plans for an experimental and computational study on the structure and extinction of low strain rate diffusion flames by a suppressant added to the oxidizer stream. Stable low strain rate flames will be established through ground based reduced gravity experiments using the 2.2 s drop tower. A variety of agents will be investigated, including both physically and chemically acting agents (He, N2, CO2, and CF3Br) for flames burning methane and propane. A computational model of flame structure and extinction will be modified to include radiative losses, which is thought to be a significant heat loss mechanism at low strain rates.

Hamins, A.↗

NASA NTRS2015

An Approach to Quantify Workload in a System of Agents

The role of humans in aviation and other domains continues to shift from manual control to automation monitoring. Studies have found that humans are often poorly suited for monitoring roles, and workload can easily spike in off-nominal situations. Current workload measurement tools, like NASA TLX, use human operators to assess their own workload after using a prototype system. Such measures are used late in the design process and can result in ex- pensive alterations when problems are discovered. Our goal in this work is to provide a quantitative workload measure for use early in the design process. We leverage research in human cognition to de ne metrics that can measure workload on belief-desire-intentions based multi-agent systems. These measures can alert designers to potential workload issues early in design. We demonstrate the utility of our approach by characterizing quantitative differences in the workload for a single pilot operations model compared to a traditional two pilot model.

Single Pilot Operations↗

NASA NTRS2000

A Unified Approach to Model-Based Planning and Execution

Writing autonomous software is complex, requiring the coordination of functionally and technologically diverse software modules. System and mission engineers must rely on specialists familiar with the different software modules to translate requirements into application software. Also, each module often encodes the same requirement in different forms. The results are high costs and reduced reliability due to the difficulty of tracking discrepancies in these encodings. In this paper we describe a unified approach to planning and execution that we believe provides a unified representational and computational framework for an autonomous agent. We identify the four main components whose interplay provides the basis for the agent's autonomous behavior: the domain model, the plan database, the plan running module, and the planner modules. This representational and problem solving approach can be applied at all levels of the architecture of a complex agent, such as Remote Agent. In the rest of the paper we briefly describe the Remote Agent architecture. The new agent architecture proposed here aims at achieving the full Remote Agent functionality. We then give the fundamental ideas behind the new agent architecture and point out some implication of the structure of the architecture, mainly in the area of reactivity and interaction between reactive and deliberative decision making. We conclude with related work and current status.

Muscettola, Nicola↗

DOE OSTI2025

Privacy-Aware RAG-Enabled LLMs for Collaborative AI in Organizations

Recent advancements in Large Language Models (LLMs) based on Transformer architectures have significantly improved capabilities in natural language processing and generation. However, deploying LLMs for inter-organizational communication poses challenges, in ensuring privacy and facilitating effective collaboration. This paper introduces a novel decentralized inference meta-agent chatbot that leverages privacy-aware Retrieval-Augmented Generation (RAG)-enabled LLMs for collaborative AI communication across organizations. Built on Microsoft’s Autogen, the platform enables LLMs to autonomously refine responses, enhancing accuracy and relevance. It incorporates advanced hallucination mitigation techniques using Uptrain and a privacy-focused RAG framework that employs synthetic document generation to protect sensitive information. Comprehensive evaluations demonstrate the platform’s effectiveness in maintaining contextual relevance and stringent privacy standards, effectively addressing critical challenges in LLM-enhanced collaborative AI communication. This work represents a significant step toward secure and efficient inter-organizational collaboration using advanced generative AI technologies.

97 - MATHEMATICS AND COMPUTING↗

DOE OSTI2026

MOSAIC-CONUS: A Multimodal, Multi-Temporally Paired Dataset for Earth Sciences

Earth embeddings—vector representations of geographic locations indexed in space and time—are emerging as a unifying interface for geospatial AI. However, their quality depends not only on model design, but on how multimodal Earth observation (EO) data are spatially indexed, temporally aligned, and cross-modally associated during pretraining. We introduce MOSAIC-CONUS (Multimodal Observations with Spatially Aligned Imagery, Urban Points of Interest, In-Situ Measurements and Text Captions), a large-scale EO dataset over the contiguous United States, organized around 250,000 stratified point indices that serve as stable spatial keys across seven modalities: active radar, passive optical imagery, lidar-derived elevation, land cover, functional context, hydrometeorological measurements, and textual summaries. Unlike existing EO datasets, MOSAIC-CONUS introduces four contributions not jointly addressed in prior work: 1. an open-source, large-scale multimodal EO corpus structured around point-indexed data designed to support Earth embedding learning; 2. explicit radar-optical pairing tables spanning twelve temporal alignment regimes, formalizing cross-sensor alignment as a controllable variable for analyzing how temporal mismatch across modalities influences learned embeddings quality; 3. a benchmark suite spanning cross-modal retrieval, annual nightlights regression, and basin-held-out streamflow prediction, positioning MOSAIC-CONUS as a benchmark-ready resource for multimodal AI systems; and 4. a language-based embedding layer through co-registered textual summaries, enabling Earth embeddings to function as a queryable interface for agentic AI systems. The dataset and pairing protocols are publicly released.

54 ENVIRONMENTAL SCIENCES↗

NASA NTRS2006

Agent-Based Simulations of Malaria Transmissions with Applications to a Study Site in Thailand

The dynamics of malaria transmission are driven by environmental, biotic and socioeconomic factors. Because of the geographic dependency of these factors and the complex interactions among them, it is difficult to generalize the key factors that perpetuate or intensify malaria transmission. Methods: Discrete event simulations were used for modeling the detailed interactions among the vector life cycle, sporogonic cycle and human infection cycle, under the explicit influences of selected extrinsic and intrinsic factors. Meteorological and environmental parameters may be derived from satellite data. The output of the model includes the individual infection status and the quantities normally observed in field studies, such as mosquito biting rates, sporozoite infection rates, gametocyte prevalence and incidence. Results were compared with mosquito vector and human malaria data acquired over 4.5 years (June 1999 - January 2004) in Kong Mong Tha, a remote village in Kanchanaburi Province, western Thailand. Results: Three years of transmissions of vivax and falciparum malaria were simulated for a hypothetical hamlet with approximately 1/7 of the study site population. The model generated results for a number of scenarios, including applications of larvicide and insecticide, asymptomatic cases receiving or not receiving treatment, blocking malaria transmission in mosquito vectors, and increasing the density of farm (host) animals in the hamlet. Transmission characteristics and trends in the simulated results are comparable to actual data collected at the study site.

Kiang, Richard K.↗

DOE OSTI2026

Powertrain Autoresearch (TETA Autoresearch: RouteE Powertrain Experiments) [SWR-26-090]

Powertrain Autoresearch is an experiment repository based on the template set in teta-autoresearch containing branch history of LLM workers investigating powertrain ML model improvements. This repository contains an instance of teta-autoresearch applied to ML model exploration for RouteE Powertrain. An AI agent and human-in-the-loop iteratively explored feature space, ML model type, and hyperparameter spaces in an attempt to improve over existing models. The repo catalogs the result for review at USRSE 2026 Conference. Related: https://github.com/NatLabRockies/teta-autoresearch.

Reinicke, Nicholas [National Laboratory of the Roc↗

NASA NTRSDate not supplied

APIS: Honeybee Foraging Task Assignment for Use in Uncertain and Unreliable Environments

Multiagent Cyber-Physical-Human (CPH) systems in realistic environments operate under uncertain conditions. Communication among agents, aimed at reducing the uncertainty, is itself subject to uncertainty. We propose to manage uncertainties in autonomous, long-duration operations of multiagent systems via a modified Honeybee Foraging (HBF) behavioral scheme. The resulting system, Autonomous Persistent Intelligent Swarm (APIS),incorporates two new behaviors to ameliorate informational uncertainty. When “scouting”, agents are tasked based on informational quality and reliability rather than solely on priorities. When “dancing”, agents are tasked to rendezvous with other dancing agents to exchange information at close range, where successful communication is guaranteed. When coupled with uncertainty-aware modeling across agents, these behaviors improve situational awareness and resilience of the system, enabling it to function under more uncertain conditions arising during long-duration missions.

Autonomous systems↗

NASA NTRSDate not supplied

APIS: Honeybee Foraging Task Assignment for Use in Uncertain and Unreliable Environments

Multiagent Cyber-Physical-Human (CPH) systems in realistic environments operate under uncertain conditions. Communication among agents, aimed at reducing the uncertainty, is itself subject to uncertainty. We propose to manage uncertainties in autonomous, long-duration operations of multiagent systems via a modified Honeybee Foraging (HBF) behavioral scheme. The resulting system, Autonomous Persistent Intelligent Swarm (APIS),incorporates two new behaviors to ameliorate informational uncertainty. When “scouting”, agents are tasked based on informational quality and reliability rather than solely on priorities. When “dancing”, agents are tasked to rendezvous with other dancing agents to exchange information at close range, where successful communication is guaranteed. When coupled with uncertainty-aware modeling across agents, these behaviors improve situational awareness and resilience of the system, enabling it to function under more uncertain conditions arising during long-duration missions.

Autonomous systems↗

DOE OSTI2025

Replace Human Intelligence with Fast and Smart Geometric Reasoning and Graph Neural Network to Accelerate Next Gen ModSim Workflows

We present an agent-guided approach to CAD geometry decomposition that automates hex/hybrid meshing with graph neural networks (GNNs) to accelerate next-generation ModSim workflows. Our end-to-end pipeline (i) reduces 3D boundary-representation (B-Rep) models to a 2D chordal axis skeleton (CAT) and then to a 1D bipartite graph of surface and curve nodes, (ii) assigns per node labels as Cubit® WebCut actions, (iii) trains a multi-action GNN under supervised learning, and (iv) predicts five surface-node and three curve-node actions on out-of-distribution test geometries. Each graph node carries geometric, topological, and meshing attributes drawn from the B-Rep “skin” and CAT “skeleton,” with two-way mappings across 3D↔2D↔1D representations to maintain traceability back to 3D CAD. The supervised learning model exhibits stable convergence of the binary cross-entropy loss and achieves 98.7% accuracy on unseen lattice models. To operationalize decision-making, we rank predicted commands by geometric significance and prototyped the agent-guided workflow through the Cubit® Meshing PowerTool GUI. As a stretch goal, we explore reinforcement learning (RL) to reduce or remove label requirements and to learn policies for action sequences that maximize total reward (e.g., size of hex-meshable regions and resulting hex mesh quality). When all-hex meshing is not feasible, the agent assists in producing hybrid meshes—prioritizing hex in critical regions and transitioning to tetrahedral elements (tets) elsewhere—maintaining fidelity while ensuring robustness. The overarching objective is to replace manual, heuristics-based decomposition with data-driven, reproducible automation, cutting meshing turnaround time by orders of magnitude. We anticipate direct impact on simulation workflows through intelligent, scalable decomposition of complex CAD models into hex-meshable subdomains.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2025

Closing the Loop between In Situ Stress Complexity and EGS Fracture Complexity

We present an agent-guided approach to CAD geometry decomposition that automates hex/hybrid meshing with graph neural networks (GNNs) to accelerate next-generation ModSim workflows. Our end-to-end pipeline (i) reduces 3D boundary-representation (B-Rep) models to a 2D chordal axis skeleton (CAT) and then to a 1D bipartite graph of surface and curve nodes, (ii) assigns per node labels as Cubit® WebCut actions, (iii) trains a multi-action GNN under supervised learning, and (iv) predicts five surface-node and three curve-node actions on out-of-distribution test geometries. Each graph node carries geometric, topological, and meshing attributes drawn from the B-Rep “skin” and CAT “skeleton,” with two-way mappings across 3D↔2D↔1D representations to maintain traceability back to 3D CAD. The supervised learning model exhibits stable convergence of the binary cross-entropy loss and achieves 98.7% accuracy on unseen lattice models. To operationalize decision-making, we rank predicted commands by geometric significance and prototyped the agent-guided workflow through the Cubit® Meshing PowerTool GUI. As a stretch goal, we explore reinforcement learning (RL) to reduce or remove label requirements and to learn policies for action sequences that maximize total reward (e.g., size of hex-meshable regions and resulting hex mesh quality). When all-hex meshing is not feasible, the agent assists in producing hybrid meshes—prioritizing hex in critical regions and transitioning to tetrahedral elements (tets) elsewhere—maintaining fidelity while ensuring robustness. The overarching objective is to replace manual, heuristics-based decomposition with data-driven, reproducible automation, cutting meshing turnaround time by orders of magnitude. We anticipate direct impact on simulation workflows through intelligent, scalable decomposition of complex CAD models into hex-meshable subdomains.

42 ENGINEERING↗

NASA NTRS1990

Identifying atmospheric monitoring needs for Space Station Freedom

The monitoring needs for Space Station Freedom were identified by examining: the experiences of past missions; ground based tests of proposed life support systems; a contaminant load model; metabolic production from an 8-person crew; and a fire scenario. Continuous monitoring is recommended for components critical for life support, and that intermittent analysis be provided for all agents that may exceed one-half the spacecraft maximum allowable concentration. The minimum monitoring effort recommended includes continuous monitoring for: N2, O2, CO2, CO, H2O, H2, CH4, nonmethane hydrocarbons, aromatic hydrocarbons, refrigerants, and halons. Information on over 70 compounds is presented on the rationale for monitoring the frequency of analysis, and concentration ranges.

Casserly, Dennis M.↗

NASA NTRS2009

Complexity for Survival of Living Systems

A logical connection between the survivability of living systems and the complexity of their behavior (equivalently, mental complexity) has been established. This connection is an important intermediate result of continuing research on mathematical models that could constitute a unified representation of the evolution of both living and non-living systems. Earlier results of this research were reported in several prior NASA Tech Briefs articles, the two most relevant being Characteristics of Dynamics of Intelligent Systems (NPO- 21037), NASA Tech Briefs, Vol. 26, No. 12 (December 2002), page 48; and Self-Supervised Dynamical Systems (NPO- 30634) NASA Tech Briefs, Vol. 27, No. 3 (March 2003), page 72. As used here, living systems is synonymous with active systems and intelligent systems. The quoted terms can signify artificial agents (e.g., suitably programmed computers) or natural biological systems ranging from single-cell organisms at one extreme to the whole of human society at the other extreme. One of the requirements that must be satisfied in mathematical modeling of living systems is reconciliation of evolution of life with the second law of thermodynamics. In the approach followed in this research, this reconciliation is effected by means of a model, inspired partly by quantum mechanics, in which the quantum potential is replaced with an information potential. The model captures the most fundamental property of life - the ability to evolve from disorder to order without any external interference. The model incorporates the equations of classical dynamics, including Newton s equations of motion and equations for random components caused by uncertainties in initial conditions and by Langevin forces. The equations of classical dynamics are coupled with corresponding Liouville or Fokker-Planck equations that describe the evolutions of probability densities that represent the uncertainties. The coupling is effected by fictitious information-based forces that are gradients of the information potential, which, in turn, is a function of the probability densities. The probability densities are associated with mental images both self-image and nonself images (images of external objects that can include other agents). The evolution of the probability densities represents mental dynamics. Then the interaction between the physical and metal aspects of behavior is implemented by feedback from mental to motor dynamics, as represented by the aforementioned fictitious forces. The interaction of a system with its self and nonself images affords unlimited capacity for increase of complexity. There is a biological basis for this model of mental dynamics in the discovery of mirror neurons that learn by imitation. The levels of complexity attained by use of this model match those observed in living systems. To establish a mechanism for increasing the complexity of dynamics of an active system, the model enables exploitation of a chain of reflections exemplified by questions of the form, "What do you think that I think that you think...?" Mathematically, each level of reflection is represented in the form of an attractor performing the corresponding level of abstraction with more details removed from higher levels. The model can be used to describe the behaviors, not only of biological systems, but also of ecological, social, and economics ones.

Zak, Michail↗

DOE OSTI2025

Holistic Microstructure Control Strategies in Photopolymerization‐Induced Phase Separation of Acrylate Systems

Open porous materials, known for their large surface area and interconnected structures, are essential in various applications, including batteries, ion exchange, catalysis, filtration, and electronic waste recycling. A critical aspect of the functionality of porous membranes is the precise control of pore size and morphology. Photopolymerization-induced phase separation (photo-PIPS) offers a convenient and versatile methods for creating porous structures. However, controlling the porous morphology remains challenging due to the complex interplay between thermodynamics, polymerization kinetics, and monomer structures, which makes it difficult to establish the relationship between processing conditions and resulting morphology in photo-PIPS. Herein, a physics-based phase-field model capable of generating and characterizing the microstructures of porous materials based on both average and localized features is developed. Using the phase-field simulations as test bed, the effects of polarity, light intensity, and curing temperature, as well as the previously unexplored roles of chain transfer agents and substrates, on the morphology of the resulting porous microstructure are investigated. Experiments are performed to verify the results predicted by the simulations. This work lays out a comprehensive guide for designing PIPS-derived porous microstructures and offers practical engineering strategies for tailoring microstructure-level topology and size of pores for application-specific needs.

36 MATERIALS SCIENCE↗

NASA NTRS1989

Cooperating intelligent systems

Some of the issues connected to the development of a bureaucratic system are discussed. Emphasis is on a layer multiagent approach to distributed artificial intelligence (DAI). The division of labor in a bureaucracy is considered. The bureaucratic model seems to be a fertile model for further examination since it allows for the growth and change of system components and system protocols and rules. The first part of implementing the system would be the construction of a frame based reasoner and the appropriate B-agents and E-agents. The agents themselves should act as objects and the E-objects in particular should have the capability of taking on a different role. No effort was made to address the problems of automated failure recovery, problem decomposition, or implementation. Instead what has been achieved is a framework that can be developed in several distinct ways, and which provides a core set of metaphors and issues for further research.

Rochowiak, Daniel↗