Evolving Classification Rules by Unconstrained Gene Expression Programming

Unconstrained Gene Expression Programming (UGEP), a new unconstrained linear encoded Gene Expression Programming (GEP), is introduced and applied to solve classification problems in this paper. Different from GEP, both amount and length of the genes are dynamically adjusted in the UGEP chromosome during the evolution process. Experiment results indicate that UGEP perform better than GEP in classification problems.