An Improved GA with Random Gene Crossover and Polyploidy Strategy

Abstract

A kind of improved genetic algorithm with real-coded crossover on random gene and polyploidy strategy (RGPGA)was proposed. Based on the term of polyploidy, The paper adopted polyploidy which composed of optimal chromatid, reserved chromatid and mutant chromatid, and real-coding to improve the diversity of population; the roulette strategy was employed to the selection operation; and a kind of crossover on random gene was proposed. At part 4, RGPGA was applied to solve benchmark problems, and compared with the standard genetic algorithm. The results show that RGPGA not only keeps the diversity of population, represses the premature convergence effectively, but also reduces the algorithm complexity, improves the precision, which can find global optima of the complex high-dimension functions with higher precision.