Adaptive Multi-Objective Particle Swarm Optimization Using Three-Stage Strategy

Abstract

It is one of the crucial problems in solving multi-objective problems (MOPs) that balance the convergence and diversity of the algorithm to obtain an outstanding Pareto optimal solution set. In order to elevate the performance further and improve the optimization efficiency of multi-objective particle swarm optimization (MOPSO), a novel adaptive MOPSO using a three-stage strategy (tssAMOPSO) is proposed in this paper, which can effectively balance the exploration and exploitation of the population and facilitate the convergence and diversity of MOPSO. Firstly, an adaptive flight parameter adjustment, formulated by the convergence contribution of nondominated solutions, can ameliorate the convergence and diversity of the algorithm enormously. Secondly, the population carries out the three-stage strategy of optimization in each iteration, namely adaptive optimization, decomposition, and Gaussian attenuation mutation. The three-stage strategy remarkably promotes the diversity and efficiency of the optimization process. Moreover, the convergence of three-stage optimization strategy is analyzed. Then, memory interval is equipped with particles to record the recent positions, which vastly improves the reliability of personal best selection. In the maintenance of external archive, the proposed fusion index can enhance the quality of nondominated solutions directly. Finally, comparative experiments are designed by a series of benchmark instances to verify the performance of tssAMOPSO. Experimental results show that the proposed algorithm achieves admirable performance compared with other contrast algorithms.

Multi-objective Optimal Scheduling Analysis of Power System Based on Improved Particle Swarm Algorithm

“…Next, differential evolution is used in the sub-problems' technique section to gain the most fulfilling scheduling scheme. In the literature [5], a multi-objective stochastic black gap particle swarm optimization algorithm is proposed to resolve the EED optimization hassle by proposing a Pareto dominance circumstance with equation constraints, so that the possible area of the answer consists of the set of Pareto highest quality solutions, from which the compromise most fulfilling answer is selected. In the literature [6], a differential evolution-crossover particle swarm algorithm is proposed to replace the crossover chance through parameter adaptive control, so that its convergence pace is higher than different algorithms.…”