Learning of Cooperative action in multiagent system : a case study of pass play in Soccer
Hitoshi Matsubara · National Conference on Artificial Intelligence · 1996
From the standpoint of multi-agent systems, soccer (association football), which is just one of usual team sports, make a good example of problems in the real world which is moderately abstracted. We have chosen soccer as one of standard problems for the study on multi-agent systems, and we are now developing a soccer server, which provides a common test-bench to various multi-agent systems. We present an experiment on cooperative action learning among soccerplayer agents on the server. Our soccer agent can learn whether he should shoot a ball or pass it. Introduction Soccer (association football) is a typical team game, in which each player is required to play cooperatively. And soccer is a real-time game in which situation changes dynamically. We have chosen soccer as one of standard problems for the study on multi-agent systems, and we are now developing a soccer server for the research. We present an experiment on cooperative action learning among soccer-player agents on the server. Soccer as a Standard Problem From the standpoint of multi-agent systems, soccer (association football), which is just one of usual team sports, make a good example of problems in the real world which is moderately abstracted. Multi-agent systems provides us with research subjects such as cooperation protocol by distributed control and effective communication, while having advantages as follows: efficiency of cooperation adaptation robustness