Multiobjective particle swarm optimization based on preference ordering optimality criterion

In the multiobjective particle swarm optimizaiton, Pareto dominance-based ranking procedures optimization become ineffective in sorting out the quality of solutions when the number of objectives is large. This effects on multiobjective particle swarm optimization because it searches optimal region according to the personal best and global best. In this paper, the concept of preference ordering is introduced as a new optimality criterion to research the high dimensional multiobjective particle swarm optimization. At the same time, the equation of velocity updating is improved according to the features of sharing information in particle swarm optimizaiton. The experiments show that new optimality criterion is effective to sorting out the solutions when the number of objectives is very large, and which can find the best compromise solutions. Finally, the performance of convergence and diversity of algorithm is improved.