Arithmetic dynamical genetic programming in the XCSF Learning Classifier System

This paper presents results from an investigation into using a continuous-valued dynamical system representation within the XCSF Learning Classifier System. In particular, dynamical arithmetic genetic networks are used to represent the traditional condition-action production system rules. It is shown possible to use self-adaptive, open-ended evolution to design an ensemble of such dynamical systems within XCSF. The results presented herein show that the collective emergent behaviour of the evolved systems exhibits competitive performance with those previously reported on a non-linear continuous-valued reinforcement learning problem. In addition, the introduced system is shown to provide superior approximations to a number of composite polynomial regression tasks when compared with conventional tree-based genetic programming.