An improved differential evolution with information intercrossing and sharing mechanism for numerical optimization

This paper presents a novel differential evolution algorithm by designing a stochastic mixed mutation strategy and an information intercrossing and sharing mechanism. To effectively avoid the premature convergence and enhance the information dissemination between subpopulations under the work specialization, a stochastic mixed mutation strategy is first proposed by incorporating a cosine perturbation into the probability parameter setting and using two mutation strategies with this probability to balance the exploration and exploitation. Then, an information intercrossing and sharing mechanism is developed to make good use of the information of individuals by dividing the population into superior and inferior subpopulations according to their fitness values and exchanging or sharing their information. Furthermore, a simple and efficient approach is applied to adjust control parameters. Finally, the proposed algorithm is compared with twelve typical algorithms by numerical experiments on 55 benchmark functions from both CEC2005 and CEC2014, and is applied to solve the Parameter Estimation for Frequency-Modulated Sound Waves. Experimental results show that the proposed algorithm is very competitive.