Learning classifier systems

Martin V. Butz

Abstract: This tutorial gives an introduction to Learning Classifier Systems focusing on the Michigan-Style type and XCS in particular. The objective is to introduce (1) where LCSs come from, (2) how LCSs generally work, (3) which different systems exist, (4) how the XCS system works, (5) how an LCS should be applied to a problem at hand, and (6) which current promising research directions and application areas are out there.