Investigating the Dynamics of Prisoner’s Dilemma Models with Hidden Markov Models

Abstract

The article proposes a methodology that can be applied to study complex dynamical systems consisting of a large number of interacting elements (components) in the case when their interactions are essentially prone to random factors. The key issue in the study of a dynamical system is the availability of an efficient method to describe its dynamics. The article proposes an approach to the description of aggregated dynamics of systems consisting of identical interacting elements. The approach is demonstrated in application to an iterated N-player Prisoner's Dilemma game. The proposed approach and the methodology of hidden Markov models are used to examine two applied problems involving classification of partially observable trajectories of the model.