Evolutionary hypergame dynamics

Abstract

A common assumption employed in most previous works on evolutionary game dynamics is that every individual player has full knowledge about and full access to the complete set of available strategies. In realistic social, economical, and political systems, diversity in the knowledge, experience, and background among the individuals can be expected. Games in which the players do not have an identical strategy set are hypergames. Studies of hypergame dynamics have been scarce, especially those on networks. We investigate evolutionary hypergame dynamics on regular lattices using a prototypical model of three available strategies, in which the strategy set of each player contains two of the three strategies. Our computations reveal that more complex dynamical phases emerge from the system than those from the traditional evolutionary game dynamics with full knowledge of the complete set of available strategies, which include single-strategy absorption phases, a cyclic competition (`rock-paper-scissors') type of phase, and an uncertain phase in which the dominant strategy adopted by the population is unpredictable. Exploiting the pair interaction and mean field approximations, we obtain a qualitative understanding of the emergence of the single strategy and uncertain phases. We find the striking phenomenon of strategy revival associated with the cyclic competition phase and provide a qualitative explanation.Our work demonstrates that the diversity in the individuals' strategy set can play an important role in the evolution of strategy distribution in the system. From the point of view of control, the emergence of the complex phases offers the possibility for harnessing evolutionary game dynamics through small changes in individuals' probability of strategy adoption.