A pursuit simulation using a time-dependent classifier system; TCS

This paper describes the time-dependent classifier system (TCS) which can obtain a better performance than a conventional classifier system in the pursuit problem. TCS handles time sequential state transitions of the environment, so it can get more detailed actions than a conventional classifier system. To realize TCS, we take notice of rules, and propose the new definition of them to express the time sequential information. We carry out several experiments using the method, and show the result of them to show how well TCS works.