Co-evolutionary classifier systems for multi-agent simulation

In this paper, MAZCS-a multi-agent system that learns using ZCS-is used for social modelling on the "El Farol" problem. Experiments with ten agents and different goal settings for the problem show that MAZCS is always able to solve it and emergent behaviour is derived from the autonomous control of each agent. The results are divided into macro and micro-level analysis, the former providing the overall performance of the system, the latter the detailed ZCS rule evolution and explanation of the agent\´s actions. Performance analysis of the rules shows that the emergent behaviour is caused by a combination of the reward received by each ZCS and its reinforcement mechanism. For its performance and action explanation MAZCS has proved to be a good modelling tool for social simulation. Furthermore the system solved the problem when using a hundred agents, assuring scalability for bigger simulation needs