A Study of Compulsive Genetic Algorithm and Its Performance

The convergence and local research ability of genetic algorithm is a well concerned research field in recent years. New evolution law of species is introduced in the paper, and based on the new evolution law, compulsive operator was introduced and a new genetic algorithm - compulsive genetic algorithm (CGA) was proposed to improve the convergence of GA. CGA takes advantage of the fitness of current and past generations to create an approximation of the evolution process, identify the evolution direction and improve their evolution progress, which will accelerate the convergence of GA. Two experimental examples were computed to test the convergence and local research ability of CGA. The experimental results show that CGA is of good convergences and good local research ability.