Towards Understanding Evolutionary Bilevel Multi-Objective Optimization Algorithm

A number of studies can be found in the context of bilevel single objective optimization problems, but not many exist, which tackle the bilevel multi-objective problems. Deb and Sinha (October, 2008) proposed a bilevel multi-objective optimization algorithm based on evolutionary multi-objective optimization (EMO) principles and discussed the issues involved in solving such a problem. In this paper we suggest an improved version of the previous algorithm which leads to high number of savings in function evaluations. Simulation results have been presented for two test problems and a comparison with the previous version has been done. The paper also discusses the complexity of bilevel problems and challenges involved in handling such problems. The existence of these problems in many practical problem solving tasks like optimal control, process optimization, game-playing strategy development, transportation problems, and others make it an important area which still needs to be considered by researchers. A two level optimization task involved in solving such problems makes the problem difficult and poses a number of challenges in getting close to the pareto front which has been addressed in the paper.