抄録
Genetic Programming (GP) is one of evolutionary computation methods and GP can handle structural data such as tree structure, and graph structure. One of the problems of GP is surplus chromosome growth which is called a bloat. In this paper, searching mechanism of GP is discussed. Through the numerical examples of two types of Symbolic Regression problems, the performance of crossover and influence of mutation rate were researched. In these researches, the influences of tree depth and mutation rate on search solution were discussed. As result, it was found out that GP doesn't search interpolation by crossover in complex problem.
