基于拥挤距离排序的多目标粒子群优化算法及其应用

Multi-objective particle swarm optimization algorithm based on crowding distance sorting and its application

针对多目标粒子群算法在全局寻优能力和Pareto集多样性上的不足,提出基于拥挤距离排序的多目标粒子群算法。该算法采用精英策略,基于个体拥挤距离降序排列,进行外部种群的缩减和全局最优值的更新,并在内部粒子群中引入小概率变异机制,增强算法的全局寻优能力,控制Pareto最优解的数目,同时保证其收敛性和多样性特征。在电梯曳引性能的多目标优化应用中,证明了该算法对于两目标和三目标优化问题求解的有效性。不同规模实例的运算对比表明,该算法在Pareto前沿的收敛性和多样性方面均优于改进强度Pareto进化算法,且缩短了运算时间,具有较高的效率与鲁棒性。

Aiming at shortcomings in global searching capacity and diversified Pareto set existing in the traditional multi-objective particle swarm optimization algorithms, a multi-objective particle swarm optimization algorithm based on crowding distance sorting was proposed. With the elitism strategy, the shrink of the external population and update of the global optimum were achieved based on individuals＇ crowding distance sorting in descending order. A small ratio mutation was introduced to the inner swarm to enhance the global searching capacity of the algorithm. And the number of Pareto optimal solutions could be controlled, the convergence and diversity of Pareto optimal set could be guaranteed as well. Effectiveness of the algorithm with two or three objectives was proved by the optimiza- tion of elevator traction performances. Comparison results among cases with different scales illustrated that this al- gorithm outperformed Strength Pareto Evolutionary Algorithm 2 （SPEA2） in the convergence and diversity characteristics of Pareto optimal front with shorter computation time, higher efficiency and robustness.