Multifactorial Differential Evolution Algorithm with Intermediate Population

Multifactorial optimization algorithm has been widely concerned because of its strong ability to solve multi-tasking optimization problems. However, the existing multifactorial evolutionary algorithms still have the problem of lack of transfer way, which will lead to negative transfer phenomenon and reduce the performance of the algorithm. In order to solve this issue, this article proposes multifactorial differential evolution algorithm with intermediate population (MFDE-IP). The intermediate population is obtained by linear weighting of optimal individuals for different tasks. In addition, a new mutation strategy is designed to guide the transfer of the population. The experimental data show that the proposed algorithm has higher convergence accuracy and faster convergence speed.