Acquisition of the various coordinated motions of multi-agent system on soccer game

This study theoretically concerns the exploitation of the possibility of acquiring the most suitable strategies for the each agent in multi-agent systems under a dynamically changing environment. A soccer game is employed as an example. The expected and acquired suitable strategies could be respected as the coordinated motions within agents and then evolve by themselves through the experiences of playing games. Here based on the stochastic learning automata, the agent MATEUS is proposed and the soccer games are executed by MATEUS. Results show that expected co-evolutionary strategies are generated