A stochastic variant of replicator dynamics in zero-sum games and its invariant measures

DOI10.1016/J.PHYSD.2023.133940zbMATH Open1530.91021arXiv2302.06969OpenAlexW4387378394MaRDI QIDQ6090682FDOQ6090682

Authors: Maximilian Engel, Georgios Piliouras 

Publication date: 17 November 2023

Published in: Physica D (Search for Journal in Brave)

Abstract: We study the behavior of a stochastic variant of replicator dynamics in two-agent zero-sum games. We characterize the statistics of such systems by their invariant measures which can be shown to be entirely supported on the boundary of the space of mixed strategies. Depending on the noise strength we can furthermore characterize these invariant measures by finding accumulation of mass at specific parts of the boundary. In particular, regardless of the magnitude of noise, we show that any invariant probability measure is a convex combination of Dirac measures on pure strategy profiles, which correspond to vertices/corners of the agents' simplices. Thus, in the presence of stochastic perturbations, even in the most classic zero-sum settings, such as Matching Pennies, we observe a stark disagreement between the axiomatic prediction of Nash equilibrium and the evolutionary emergent behavior derived by an assumption of stochastically adaptive, learning agents.

Full work available at URL: https://arxiv.org/abs/2302.06969

Recommendations

• Best-response dynamics for evolutionary stochastic games
• Stochastic evolutionary game dynamics: foundations, deterministic approximation, and equilibrium selection
• The emergence of rational behavior in the presence of stochastic perturbations
• Stable and historic behavior in replicator equations generated by similar-order preserving mappings
• Stochastic replicator dynamics: a theoretical analysis and an experimental assessment

zbMATH Keywords

game theoryinvariant measuresNash equilibriumstochastic differential equationsreplicator equationsgenerative adversarial networks

Mathematics Subject Classification ID

Noncooperative games (91A10) 2-person games (91A05) Stochastic games, stochastic differential games (91A15)

Cites Work

• Title not available (Why is that?)
• Prediction, Learning, and Games
• Algorithmic Game Theory
• Title not available (Why is that?)
• Title not available (Why is that?)
• Utility Theory without the Completeness Axiom
• Evolutionary dynamics with aggregate shocks
• The long-run behavior of the stochastic replicator dynamics
• Evolutionary dynamics for bimatrix games: A Hamiltonian system?
• Online learning and online convex optimization
• Title not available (Why is that?)
• Persistence in fluctuating environments
• Time averages, recurrence and transience in the stochastic replicator dynamics
• Exponential weight algorithm in continuous time
• The emergence of rational behavior in the presence of stochastic perturbations
• Adaptive game playing using multiplicative weights
• The multiplicative weights update method: a meta-algorithm and applications
• An iterative method of solving a game
• Dynamics of games and genes: Discrete versus continuous time
• Robust permanence and impermanence for stochastic replicator dynamics
• Stochastic Approximations and Differential Inclusions, Part II: Applications
• Title not available (Why is that?)
• Title not available (Why is that?)
• A general theory of coexistence and extinction for stochastic ecological communities
• Random switching between vector fields having a common zero
• Cycles in adversarial regularized learning
• Optimization Despite Chaos: Convex Relaxations to Complex Limit Sets via Poincaré Recurrence
• Coexistence and extinction for stochastic Kolmogorov systems
• The stabilization of equilibria in evolutionary game dynamics through mutation: mutation limits in evolutionary games
• A classification of the dynamics of three-dimensional stochastic ecological systems

Cited In (2)

• Title not available (Why is that?)
• Global Convergence for Replicator Dynamics of Repeated Snowdrift Games

This page was built for publication: A stochastic variant of replicator dynamics in zero-sum games and its invariant measures

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6090682)