Multi-robot cooperation and competition with genetic programming

Abstract

In this paper, we apply Genetic Programming(GP) on multi-robot cooperation and competition problem. GP is taken as a real time planning method in stead of learning method. Robot all use GP to make a plan and then walk according to the plan. The environment is composed of two parts, natural environment, which is the obstacles, and social environment that refers to other robots. The cooperation process is accomplished by robot’s adaptation to both of them. In spite of the fact that there is no communication among robots and little knowledge about how to cooperate well, the adaptive capability in dynamic environment enable robots to complete a common task or solve the competition. Several experiments are taken and the results are shown.