Coevolução de agentes de estado finito / Coevolution of finite state agents
AUTOR(ES)
Marcus Vinicius Pereira Pessôa
DATA DE PUBLICAÇÃO
1999
RESUMO
This work is concerned on the study of evolutionary algorithms, in the solution of problems involving coevolution of agents, whose behavior correspond to finite state machines. The genetic algorithms are the base for the coevolutive process, approaching cooperative and competitive behaviors. The cooperation happens inside of a specific group of agents and the competition happens among these groups. For this study were varied several genetic algorithms characteristics and the ambient where the interactions among the agents happen. Factors as the representation and control of the population s elements operated by the genetic algorithm, its selection and operation were explored, in way to evaluate which outlines would be more convenient for application in those types of problems. For the attainment of this work, the Swarm simulation system, developed by Santa Fe Institute - USA, was used.
ASSUNTO(S)
teoria de jogos cooperation cooperação game theory competição genetic algorithms artificial intelligence competition inteligência artificial algoritmos genéticos
