Using Cognitive Behavioral Learning in Multi-agent Pursuit-Evasion Game

Abstract

It is a challenging problem that Multi-Agent systemmake the best response action in a dynamic environment. This study applied probabilistic theory and Piaget's schema theory to develop a cognitive behavioral learning mechanism for the pursuit-evasion game in a dynamic environment. The proposed cognitive schema is divided into two parts, one is a perceptional schema and the other is an intentional schema. According to the agent's state, intentional schema applied probabilistic theory to propose three strategies for prediction of the evader's position in a dynamic environment. Perceptional schema applied case base dreasoning to select the adequate action for the pursuitagent. When an agent faces a new situation, it might useassimilation mechanism to deal with the problem. If the currentcognitive schema cannot explain their environment, accommodation mechanism is used to adapt the cognitive schema of the agent for dealing with their problem.