Stephanie Forrest, John Miller
Paper #: 89-005
Classifier systems are increasingly being applied to the analysis of economic phenomena. Among these applications are adaptive models of learning, the creation of artificial economies, and the development of economic webs. A methodology is described for studying the dynamical behavior of classifier systems. The methodology is useful because of the current lack of analytical results describing interactions among the various components of classifier systems. A mapping is defined between classifier systems and an equivalent dynamical system (Boolean networks). The mapping provides a way to understand and predict classifier system behaviors by observing the dynamical behavior of the Boolean networks. The paper reports initial results produced by the methodology and discusses the implications of this approach for classifier systems.