A multi-objective evolutionary algorithm based on membrane systems

In this paper, a novel algorithm based on membrane systems theory is proposed for solving multi-objective optimization problems. The proposed algorithm is composed mainly by some elements, such as symbol-objects, multi-sets, regions, rules and so on, which are inspired by the structure and functioning of the living cell. In the inner regions of membrane systems, a symbol-object represents a candidate solution of the multi-objective optimization problem. Thereafter, some symbol-objects may construct a multi-set which is sent into the appointed region by the specific rule. In addition, some evolution rules are employed to evolve the multi-set in the inner region of elementary. Moreover, the diversity of the proposed algorithm is improved by the maintaining strategy and the rules of division and dissolution. Comparative study shows that the proposed method performs better in four performance metrics when solving these benchmark problems compared with three the state-of-art algorithms. Sensitivity analysis indicates that it could quickly obtain the approximate Pareto front and satisfy the requirement of diversity of Pareto front. So, it is feasible and effective to solve multi-objective optimization problems.