Multipopulation genetic algorithm with adaptive search area

To solve the problem of slow convergence speed of the standard genetic algorithm (SGA), the strategy of adaptively changing the search area is used to reduce the search area progressively in this paper. The tactics of concerted evolution among multiple populations is proposed aimed at the deficiency of easily plunging into a local optimal solution of SGA. Distant hybridization strategy and a new method of adaptively changing crossover rate are presented by combining the scheme of multi-population and the thought of elitist population. Considered the different range of decision variables, the new definitions of individual distance and population distance are put forward, which avoids the false distance caused by traditional hamming distance. These methods can not only ensure the independence of subpopulations, but also strengthen their cooperation, improve the use ratio of excellent genes, and enhance the global search ability of GA. Finally, the effectiveness of the proposed algorithm was verified by three typical testing functions.