Evolution of memory strategies in alternating stochastic games

Authors:

Abstract
Stochastic games provide a powerful theoretical framework for capturing the feedback between individual behavior and environmental states, yet most existing studies assume that players interact synchronously. Here, we investigate the evolution of classical pure memory-one strategies in alternating stochastic games, in which players take turns making decisions based on historical outcomes and subsequently altering the environmental state. By analyzing the equilibrium conditions of pure memory-one strategies, we show through theoretical and numerical results that environmental feedback can substantially promote cooperation under alternating interactions. In particular, sustaining cooperation relies primarily on the firm-but-fair (FBF) strategy rather than the celebrated win-stay, lose-shift (WSLS) strategy. Moreover, when the environmental transition mechanism incentivizes mutual cooperation by maintaining a good state and penalizes any defection by switching the environment to a bad state, the population can form a stable, multi-strategy cooperative alliance that collectively sustains cooperation. These findings remain robust under both rare and positive mutation rates, providing novel insights into how environmental feedback and sequential interactions jointly shape the evolution of cooperation.