Differential Evolution with Controlled Annihilation and Regeneration of Individuals and A Novel Mutation Scheme

Differential Evolution is a stochastic, population-based optimization algorithm, which grew out of the need to optimize real-parameter, real-valued functions. The Differential Evolution variant that we propose to describe in this paper modifies the mutation scheme of the variant DE/best/1. We propose a three tier mutation scheme, to be suitably carried out on selected sections of the population in question. Also, the proposed variant tries to lessen the myriad troubles posed by stagnation, which is a problem faced by all Differential Evolution algorithms. Our comparative studies indicate that the proposed variant is able to compete in a direction parallel to the state-of-the-art Differential Evolution variants like JADE and jDE.