PhysicsAgentABM: Physics-Guided Generative Agent-Based Modeling

TL;DR

PhysicsAgentABM introduces a scalable, calibrated agent-based modeling framework that combines symbolic, neural, and LLM components to improve simulation accuracy and efficiency across various domains.

Contribution

It presents a novel neuro-symbolic architecture and an LLM-based clustering method, reducing LLM calls and enhancing simulation calibration and scalability.

Findings

Improved event-time accuracy over baselines

Enhanced calibration of transition distributions

Reduced LLM calls by up to 6-8 times

Abstract

Large language model (LLM)-based multi-agent systems enable expressive agent reasoning but are expensive to scale and poorly calibrated for timestep-aligned state-transition simulation, while classical agent-based models (ABMs) offer interpretability but struggle to integrate rich individual-level signals and non-stationary behaviors. We propose PhysicsAgentABM, which shifts inference to behaviorally coherent agent clusters: state-specialized symbolic agents encode mechanistic transition priors, a multimodal neural transition model captures temporal and interaction dynamics, and uncertainty-aware epistemic fusion yields calibrated cluster-level transition distributions. Individual agents then stochastically realize transitions under local constraints, decoupling population inference from entity-level variability. We further introduce ANCHOR, an LLM agent-driven clustering strategy based on cross-contextual behavioral responses and a novel contrastive loss, reducing LLM calls by up to 6-8 times. Experiments across public health, finance, and social sciences show consistent gains in event-time accuracy and calibration over mechanistic, neural, and LLM baselines. By re-architecting generative ABM around population-level inference with uncertainty-aware neuro-symbolic fusion, PhysicsAgentABM establishes a new paradigm for scalable and calibrated simulation with LLMs.