基于量子行为粒子群优化算法的路径规划

Path Planning Based on Qantum-Behaved Particle Swarm Optimization Algorithm

针对移动机器人的路径规划问题,提出了一种基于QPSO算法的改进算法。在QPSO算法的随机初始化阶段,种群的多样性较高,但随着进化的推进,个体的差异性减小,粒子群的多样性降低,致使算法容易陷入局部最优而出现早熟现象。针对这一不足,利用正态云模型的随机性和稳态倾向性,引入云模型的变异操作,使进化算法的优点与量子行为粒子群算法充分结合起来,提高QPSO在路径搜索中的性能。通过QPSO算法与改进的QPSO算法的仿真实验表明云模型变异操作的引入有效地避免了种群陷入局部搜索,较大程度上提高了路径搜索的速度。

For mobile robot path planning problem,this paper proposes an improved algorithm based on the QPSO algorithm.In the random initialization phase of the QPSO algorithm,the diversity of population is higher.But along with the advancement of evolution,the differences of particles decreases and the diversity of population is reduced,which make the algorithm tend to fall into local optimum and appearing premature phenomenon.Towards this disadvantage,this paper introduces the variation operation,using randomness and steady-state orientation of normal cloud model,and makes the advantages of the evolutionary algorithm and the QPSO algorithm to combine sufficiently,in order to improve QPSO performance in the search path.The simulation results show that the introduction of variation operation of cloud model in the improved QPSO algorithm effectively avoids the population into local search,largely improves the speed of the search path compared with the QPSO algorithm.