Grammar Model-based Program Evolution

In Evolutionary Computation, genetic operators, such as, mutation and crossover, are employed to variate the individuals to generate next population. However, these fixed, problem independent genetic operators may destroy the subsolution, usually called building blocks, instead of discovering and preserving them. One way to overcome this problem is to build a model based on the good individuals and sample this model to obtain the next population. There some research Because of the complexity of Genetic Programming (GP) tree representation, little work of this kind has been done in GP. In this paper, we propose a new method, Grammar Model-based Program Evolution (GMPE) to evolved GP program. We replace common GP genetic operator with a Probabilistic Context-free Grammar (SCFG). In each generation, a SCFG is learned and new population is generated by sampling this SCFG model. On two benchmark problem we have studied, GMPE significantly outperforms conventional GP, usually a few times faster and more reliable.