Interação como característica evolucionária em uma população modelada por autômatos celulares e algoritmo genético

Abstract

When the game theory was applied to the evolution of populations in biology, evolutionary game theory was born. Since its rst formalization with John Maynard Smith and George R. Price in 1973, an individual's strategy has been the key to population evolution. That is, the strategy of an individual in a given game was the information that evolved in the population. In this paper, we expand the characteristics of the individual who evolve, and beyond the strategy, the number of games started by the individual, as well as the maximum radius of interaction with their neighbors also evolve. Individuals and the process of population renewal are based on evolutionary characteristics of the genetic algorithm, and strategy, the number of games and interaction radius are considered as the chromosome of the individual. The population is modeled by a continuous probabilistic cellular automata, and the interactions between individuals modeled by two classic games: prisoner's dilemma and hawk-dove. Finally, two rules for updating the death-birth process are responsible for population renewal. The objective of this work was to verify how the characteristics of individuals evolve along with the population and how the population mutation rate in uenced the population evolution. Five computational experiments were performed to verify the temporal evolution, the in uence of the mutation rate on the number of games and radius of movement, the concentration of the individuals as a function of their movement and radius of operation, the distribution of the individuals on the cellular automata and the number of games and level of cooperation. We observed that the best strategic responses were in uenced by the type of game, and population renewal dynamics. In addition, the spatial distribution of individuals was also in uenced by these factors.