Exit-and-Join Dynamics for Decentralized Coalition Formation

2026-06-19 · Source: cs.MA updates on arXiv.org · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems · Depth: Expert, extended

Summary

This paper introduces a decentralized model for coalition formation, where agents make unilateral exit-and-join decisions based on local payoff comparisons using the Aumann-Drèze value. The model links cooperative payoff allocation with noncooperative best-response behavior, defining terminal partitions as coalition structures without individually profitable deviations. The research establishes equilibrium characterizations, identifies conditions for scalar Lyapunov or exact-potential representations, and analyzes how switching and acceptance costs influence local stability. Numerical experiments, including 180 independently generated games with 24-30 agents, confirm finite-time stabilization, demonstrate cost sensitivity, and test a convex-game benchmark where the grand coalition is dynamically selected.

Key takeaway

For AI Scientists designing multi-agent systems or Research Scientists modeling socio-technical dynamics, understanding these decentralized exit-and-join dynamics is crucial. Your system's emergent coalition structures will be shaped by local incentives and costs, not just global efficiency. Consider implementing Aumann-Drèze-based payoff rules and carefully tuning switching and acceptance costs to guide agents toward desired, stable configurations, recognizing that global optimality may require additional coordination mechanisms.

Key insights

Decentralized coalition formation emerges from local, Aumann-Drèze-based exit-and-join decisions, converging to stable equilibria.

Principles

• Aumann-Drèze value enables local payoff evaluation.
• Convexity ensures grand coalition efficiency, not dynamic selection.
• Marginal alignment is key for dynamic convergence.

Method

Agents evaluate exit-and-join actions by comparing current Aumann-Drèze payoffs against potential new payoffs, net of switching costs, subject to destination coalition acceptance.

In practice

• Model multi-agent AI task allocation.
• Analyze labor market mobility patterns.
• Understand subscription market churn dynamics.

Topics

• Coalition Formation
• Aumann-Drèze Value
• Multi-Agent Systems
• Game Theory
• Decentralized Dynamics
• Lyapunov Analysis

Best for: AI Scientist, Research Scientist

Related on AIssential

• Dynamic Coordination Strategy Selection for Enterprise Multi-Agent Systems — Dynamic strategy selection for enterprise multi-agent systems is a near-best routing heuristic, not a deterministic winner-selection law.
• Convex Markov Games and Beyond: New Proof of Existence, Characterization and Learning Algorithms for Nash Equilibria — Nash equilibria in General Utility Markov Games coincide with fixed points of projected pseudo-gradient dynamics.
• The Dynamics of Policy Gradient in Social Dilemmas with Partner Selection — Partner selection and population variance are critical for cooperation in multi-agent policy gradient learning.
• How Intelligence Emerges: A Minimal Theory of Dynamic Adaptive Coordination — Intelligence emerges from dynamic adaptive coordination in multi-agent systems via recursively closed feedback.
• Cattle Trade: A Multi-Agent Benchmark for LLM Bluffing, Bidding, and Bargaining — Complex multi-agent benchmarks reveal LLM strategic reasoning limitations in integrated economic games.
• Transitivity Meets Cyclicity: Explicit Preference Decomposition for Dynamic Large Language Model Alignment — Explicitly decomposing human preferences into transitive and cyclic components improves LLM alignment and robustness.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by cs.MA updates on arXiv.org.