Multi-Population Based Ensemble Mutation Method for Single Objective Bilevel Optimization Problem

Differential evolution algorithm is an effective and population-based global optimization algorithm, which has been successfully used in many different fields. In this paper, the proposed algorithm attempts to combine the advantage of the evolutionary algorithm and local search to find the global optimum solutions with the low computational cost for the single objective bilevel optimization problem. For the upper level optimization, a multi-population-based ensemble mutation method is proposed to enhance the convergence rate of the algorithm and diversity maintenance. For the lower level optimization, a local search based on sequential quadratic programming method is used to find the best solution and then return the final solution into the upper level optimization. To verify the performance of the proposed algorithm, eight benchmark functions chosen from the literature are employed. Compared with some previous evolutionary algorithms, the results show the superior performance of the proposed algorithm over other algorithms in handling single objective bilevel optimization problem.