Authors: Mengjie Zhang, Phillip Wong
Source: GZipped PostScript (54kb); Adobe PDF (154kb)
This paper describes an approach to online simplification of evolved programs in genetic programming. Rather than manually simplifying genetic programs after evolution for interpretation purpose only, this approach automatically simplifies genetic programs during the evolutionary process. In this approach, algebraic simplification rules, algebraic equivalence and prime techniques were used to simplify genetic programs. The GP system with this simplification is examined and compared to a standard GP system on a regression problem and a classification problem. The results suggest that, at certain frequencies or proportions, the GP system with this simplification can not only achieve superior performance to the standard system on these problems, but also significantly reduce the sizes of evolved programs.
Keywords: Genetic programs, program simplification, simplification proportion, simplification frequency