Evolutionary Stability in Simple Classifier Systems

In this paper, the relatively new branch of mathematics known as Evolutionary Game Theory is proposed as a potentially useful tool when seeking to resolve certain of the more global, unanswered questions related to classifier systems. In particular, it is proved that, under certain mild assumptions, the performance of a classifier systems plan will, if the Bucket Brigade Algorithm is adopted, conform to what is referred to as ‘an evolutionary stable state’. A simple example is also provided to confirm the theoretical findings.