针对现有微粒群算法在动态优化问题中容易陷入局部极值问题,提出了一种改进的动态微粒群算法——AVPSO。AVPSO用所有微粒局部最优值的平均值来代替全局最优值,通过有目的的重新初始化部分微粒扩大种群搜索范围,在感知到环境发生变化时迅速、准确地实现对目标的跟踪。实验结果表明,在求解动态优化问题时,AVPSO表现出很好的性能。将AVPSO应用于群体动画中,实现了群体路径规划的自动化。 Aiming at the problem that existing particle swarm optimization (PSO) is easy to fall into local extremum problems, an improved dynamic particle swarm optimization algorithm (AVPSO) is proposed. AVPSO replaces the global optimal value with the average value of the local optimal values ​​of all particles, and aims to rapidly and accurately track the target by purposefully reinitializing part of the particles to expand the search range of the population and perceiving changes in the environment. Experimental results show that AVPSO shows good performance in solving dynamic optimization problems. The application of AVPSO in group animation has realized the automation of group path planning.