Comparative Study on Differential Evolution: Effect of the Ratio Between Binomial Crossover and Exponential One on Results

Among directions of intelligent computing, single objective optimization with real parameter is an important research focus. In recent, setting in single objective optimization with real parameter changes dramatically. Among types of population-based metaheuristic, differential evolution performs very outstandingly. So far, among the existing three crossover strategies, the binomial and exponential crossover is rarely used. However, IMODE, a DE algorithm suitable for the latter setting, is with the binomial and exponential crossover. In this paper, we optimize the parameter of the crossover strategy by experiment. According to our results, the original value of the parameter is not the best choice.