Multi-objective Particle Swarm Optimization Based on Bipolar Preferences Control

Considering the control of bipolar preferences to the particle swarm,a method with the bipolar preferences,positive and negative preferences,is presented to lead the particle to move toward the true Pareto front.The similarities of the non-dominated solutions to bipolar preferences are computed according to TOPSIS(technique for order preference by similarity to ideal solution) decision method,then the similarities are sorted in descending order and the non-dominated solutions are maintained in the out-archives according to the order.The spread of the solutions are also determined by the similarity.Six different test functions are chosen in the simulation experiment.,and this method is compared with the unipolar preferences based MOPSO in generation distance,spacing metric and hyper-volume metric,the final comparison result show the proposed method is better in aspect of convergence and combination property.