Usage of ZCS evolutionary classifier system as a rule maker for cleaning robot task

Abstract

This paper introduces the Cleaning robot task which is a simulation of the cleaning of a room by a robot. The robot must collect all the junk in the room and put it into a container. It must take out the junk sequentially, because the amount of carried trash is limited. The actions of this robot are selected by using the Michigan style classifier system ZCS. This paper shows the capability of this system to select good rules for the robot to perform the cleaning task.