Parameter-free genetic algorithm inspired by "Disparity theory of evolution"

Abstract

We propose a novel Genetic Algorithm which we call a Parameter-free Genetic Algorithm (PfGA) inspired by the "disparity theory of evolution". The idea of the theory is based on different mutation rates in double strands of DNA. Furthermore, its idea is extended to a very compact and fast adaptive search algorithm accelerating its evolution based on the variable-size of population taking a dynamic but delicate balance between exploration (i.e., global search) and exploitation (i.e., local search). The PfGA is not only simple and robust, but also does not need to set almost all genetic parameters in advance that need to be set up in other Genetic Algorithms. To verify the effectiveness of the PfGA, we compared its results with those on the first Internatinal Contenst on Evolutionary Optimization at ICEC'96 using some recent function optimization problems. A parallel and distributed PfGA architecture is being investigated as an extension of this work, some preliminary results of which are shown.