Abstract
Genetic Algorithms (GAs) are known as adaptive heuristic search algorithms to find approximate solutions, but the problem of premature convergence and fall into local solution are remained to be solved. We had already proposed FASPGA and proved its effectiveness. However, FASPGA adopted only maximum and average fitness as the inputs of fuzzy rules, which contains not enough information to describe the search stage. In this paper, we imported two parameters (genotypic parameters and phenotype parameters) into the fuzzy rule that makes many combinations as the input of fuzzy rules. So, we performed many simulations and compared the results to find an optimum combination, which has better performance in many kinds of test functions.
