Multi agent systems improved by splitting psychological and semantic states
Scaling Multi-Agent Systems with Prospect-State Propagation
Multiagent SystemsComputers and Society
Summary
Simulating many agents at once is hard because their detailed behaviors take up too much computing power. The authors found that as simulations run, agents start to act more similarly, losing diversity. To fix this, they created a method called Prospect-State Propagation that breaks each agent’s state into two parts: a simple psychological part that keeps track of personal differences, and a detailed semantic part that uses language models to handle thinking and planning. This approach helps keep agents diverse while still allowing complex behaviors, making simulations more scalable.
multi-agent systemslarge language modelssimulationagent heterogeneityprospect theorysemantic statepsychological statestate propagationeconomic simulationscalability
Authors
Zhimei Chen, Mu Chen, Fakhri Karray
Abstract
Current LLM-based multi-agent systems (MAS) periodically compress intermediate states to reduce inference-time token consumption, thereby attempting to incorporate more agents. However, naive scaling strategies face challenges. For example, in economic simulations, large-scale MAS typically discard semantically rich economic states, i.e., agent behavioral trajectories, which are key drivers of macroeconomic fluctuations. In this paper, we reveal a phenomenon in which agent heterogeneity gradually decreases during simulation, and propose Prospect-State Propagation for Multi-Agent Systems (PspMAS). Inspired by prospect theory, PspMAS decouples each agent's micro state into a compact Prospect State and an expressive Semantic State. The former records psychological traces through a lightweight, parallelizable propagator and continuously injects heterogeneity into the system. The latter leverages the strong perception, reasoning, planning, and decision-making abilities of LLMs. These two components work complementarily, providing a scalable LLM-based multi-agent simulation solution.