Sink equilibria and the attractors of learning in games

2025-02-11

AI Summary

This paper investigates the correspondence between attractors—limit behaviors of learning dynamics (replicator dynamics and fictitious play)—and sink equilibria in game-theoretic settings. Method: Employing topological construction, preference graph analysis, and dynamical systems theory, the authors first construct counterexamples to refute the conjecture that attractors and sink equilibria are in one-to-one correspondence. They then derive necessary and sufficient conditions for this correspondence to hold. Contribution/Results: They establish that, in two-player games without weak local sources, attractors coincide exactly with sink equilibria. Moreover, they extend this framework to fictitious play for the first time, proving that its attractors are always subsets of the sink equilibrium set and are in bijective correspondence with the maximal strongly connected components of the preference graph. This work unifies the limiting structures of two canonical learning dynamics and establishes a rigorous bridge between dynamical behavior and combinatorial structure.

Technology Category

• Game Theory and Economic Paradigms: Equilibrium
• Multiagent Systems: Mechanism Design
• Knowledge Representation and Reasoning: Preferences

Application Category

• Economics, Online Markets and Human Computation: Social networks and social learning
• Web Mining and Content Analysis: Models for Web evolution
• Social Networks and Social Media: Media and governance, opinion dynamics, filter bubbles, polarization

Abstract

Characterizing the limit behavior -- that is, the attractors -- of learning dynamics is one of the most fundamental open questions in game theory. In recent work in this front, it was conjectured that the attractors of the replicator dynamic are in one-to-one correspondence with the sink equilibria of the game -- the sink strongly connected components of a game's preference graph -- , and it was established that they do stand in at least one-to-many correspondence with them. We make threefold progress on the problem of characterizing attractors. First, we show through a topological construction that the one-to-one conjecture is false. Second, we make progress on the attractor characterization problem for two-player games by establishing that the one-to-one conjecture is true in the absence of a local pattern called a weak local source -- a pattern that is absent from zero-sum games. Finally, we look -- for the first time in this context -- at fictitious play, the longest-studied learning dynamic, and examine to what extent the conjecture generalizes there. We establish that under fictitious play, sink equilibria always contain attractors (sometimes strictly), and every attractor corresponds to a strongly connected set of nodes in the preference graph.

Problem

• Characterize attractors in game theory
• Verify sink equilibria correspondence
• Examine fictitious play dynamics

Innovation

• Topological construction disproves conjecture
• One-to-one conjecture validated sans weak local source
• Fictitious play links sink equilibria to attractors