Abstract: Aiming to handle the Multi-objective Optimization Problem（MOP）, using the method of introducing multi-vector, an improved multi-objective particle swarm optimization algorithm is proposed in this paper. Improved multi-objective particle swarm optimization algorithm can find the global optimum faster than the standard multi-objective particle swarm optimization algorithm under high-dimensional situation. Through theoretical derivation, the correctness of improved multi-objective particle swarm optimization algorithm is proved. The correctness of the theoretical derivation is verified by experiment.

Key words: multi-point velocity vector, multi-objective problem, particle swarm optimization algorithm

摘要： 针对多目标粒子群算法在高维条件下易早熟、迭代步骤数较多的问题，通过引入多点速度向量，提出一种基于多点速度向量的多目标粒子群改进算法，由于改进的多目标粒子群可以看成多个对于目标函数和当前种群的多目标最优点独立的速度和位置分量的叠加，减少了在目标函数最优值搜索之间相互的影响，从而有效地提高多目标粒子群在高维条件下的收敛速度以及准确性，理论证明这这种改进的有效性。实验结果证明了理论推导的正确性。

关键词: 多点速度向量, 多目标问题, 粒子群算法