An efficient real-coded genetic algorithm for numerical optimization problems

This paper proposes an improved real-coded genetic algorithm(RCGA) with a new crossover operator and a new mutation operator. The crossover operator is designed, based on the evolutionary direction provided by two parents, the fitness ratio of two parents, and the distance between two parents. This crossover operator can improve the convergence speed of RCGAs by using the heuristic information mentioned above. Moreover, the proposed mutation operator, which utilizes the entropy information of every gene locus in chromosomes, can prevent the premature convergence of RCGAs. Experiments on benchmark test functions with different hardness describe the effectiveness of the improved RCGA.