CRGA—A Genetic Algorithm Based on Preserving Global Commonality Schemata and Restricting Local of Crossover

A new genetic algorithm based on preserving global commonality schemata and restricting local of crossover is proposed. This algorithm resolves the problem that is the two disadvantages of a standard crossover operator: disruption of high rank, long and good schemata and inefficiency between similar individuals. This algorithm protects the high rank, long and good schemata by estimating the probability of parent alleles preserved in son based on the statistic of all individuals whose fitness is better than average fitness of population, and ensures to produce new individual by restricting the local of crossover. The experiment result shows that the convergence precision and the convergence speed of this algorithm are evidently better than that of genetic algorithm based on standard crossover.