Stochasticity and evolutionary stability

Abstract

In stochastic dynamical systems, different concepts of stability can be obtained in different limits. A particularly interesting example is evolutionary game theory, which is traditionally based on infinite populations, where strict Nash equilibria correspond to stable fixed points that are always evolutionarily stable. However, in finite populations stochastic effects can drive the system away from strict Nash equilibria, which gives rise to a new concept for evolutionary stability. The conventional and the new stability concepts may apparently contradict each other leading to conflicting predictions in large yet finite populations. We show that the two concepts can be derived from the frequency dependent Moran process in different limits. Our results help to determine the appropriate stability concept in large finite populations. The general validity of our findings is demonstrated showing that the same results are valid employing vastly different co-evolutionary processes.

References in corpus (5)

• Coevolutionary Dynamics: From Finite to Infinite Populations
• Prisoners' dilemma in real-world acquaintance networks: Spikes and quasi-equilibria induced by the interplay between structure and dynamics
• The long-run behavior of the stochastic replicator dynamics
• Similarity based cooperation and spatial segregation
• Nongaussian fluctuations arising from finite populations: Exact results for the evolutionary Moran process

Cited by in corpus (2)

• Fixation times in evolutionary games under weak selection
• Noise-guided evolution within cyclical interactions