Multi-objective Particle Swarm Optimization Algorithm With Double Thresholds Based on Preference Information

This paper focused on how to obtain optimal set effectively in preference regions. A multiobjective particle swarm optimization algorithm with double thresholds( DT-MOPSO) was proposed. The algorithm used g-dominated to increase selection pressure and controled the number of non-inferior solutions by means of the beam distance threshold. The diversity threshold value was introduced to adjust the solution diversity. When the diversity value was lower than the threshold value,the adaptive grid algorithms was used to improve the diversity. Simulation results of a series of classical problems show the correctness and effectiveness of this algorithm.