Abstract: Aiming at the problem that Automated Guided Vehicle (AGV) global path planning algorithm easy to fall into local minimum and slow convergence,the traditional Elitism Tactic was enhanced with Grey Wolf Optimization (GWO) to prevent population diversity worse gradually and improve its global search capability.To prevent genes on chromosomes to cluster around a small neighborhood,an Improved Genetic Algorithm (IGA) enhanced with adaptive mutation and crossover probability based on chromosome information entropy was proposed,in which a neighborhood mutation operator was adopted for gene fragments that intersected the obstacles to be avoided quickly.An AGV path planning platform for the proposed IGA was developed by using Matlab GUI tools.Experimental results showed that the proposed IGA and the developed platform could efficiently and reliably solve complex static environment AGV path planning problems.

Key words: grey wolf optimization algorithm, information entropy, neighborhood mutation, adaptive mutation

摘要： 针对自动导引小车全局路径规划算法收敛慢和容易陷入局部最小值的问题,结合灰狼优化算法改进传统的精英保留策略,避免了传统精英保留策略使种群多样性变差的缺点,增强了全局搜索能力;为了防止染色体上的基因聚集到小的邻域内,提出了基于染色体信息熵的自适应变异和交叉概率的改进遗传算法,其中对于与障碍物相交的染色体片段采用邻域变异算子,使染色体片段快速避开障碍物。采用MATLAB GUI工具开发出基于改进遗传算法的移动机器人路径规划平台。实验结果表明,本文所提出的改进算法和开发平台能高效并可靠地求解复杂静态环境中的移动机器人路径规划问题。

关键词: 灰狼优化算法, 信息熵, 邻域变异, 自适应变异