On Symmetry for Characterizing Agent Heterogeneity in Multi-Agent Systems

DOI10.48550/ARXIV.2610.18538MaRDI QIDQ7400252arXivFDO

Authors Jianhong Wang, Jonathan Lawry

Publication date 13 October 2026

Published in arXiv

Tldr We introduce a framework for characterizing agent heterogeneity via policy-evaluation symmetries in finite-horizon partially observable stochastic games (POSGs), where symmetries form a group that classifies agents into structural types independent of policies and initial states.

Full work available at URL https://doi.org/10.48550/ARXIV.2610.18538

arXiv classification

cs.AIcs.MA

Cites work

• Causal inference in statistics: an overview
• Exploiting symmetries for single- and multi-agent partially observable stochastic domains
• How bad is selfish routing?
• Optimal and approximate Q-value functions for decentralized POMDPS
• Risk-Sensitive and Robust Decision-Making: a CVaR Optimization Approach
• Subjectivity and correlation in randomized strategies
• Symmetric games revisited
• The Complexity of Decentralized Control of Markov Decision Processes

This page was built for publication: On Symmetry for Characterizing Agent Heterogeneity in Multi-Agent Systems

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7400252)