This paper is a study of the dynamic path planning problem of the pull-type multiple Automated Guided Vehicle (multi-AGV) complex system. First, based on research status at home and abroad, the conflict types, common planning algorithms, and task scheduling methods of different AGV complex systems are compared and analyzed. After comparing the different algorithms, the Dijkstra algorithm was selected as the path planning algorithm. Secondly, a mathematical model is set up for the shortest path of the total driving path, and a general algorithm for multi-AGV collision-free path planning based on a time window is proposed. After a thorough study of the shortcomings of traditional single-car planning and conflict resolution algorithms, a time window improvement algorithm for the planning path and the solution of the path conflict covariance is established. Experiments on VC++ software showed that the improved algorithm reduces the time of path planning and improves the punctual delivery rate of tasks. Finally, the algorithm is applied to material distribution in the OSIS workshop of a C enterprise company. It can be determined that the method is feasible in the actual production and has a certain application value by the improvement of the data before and after the comparison.

本文是对拉式多自动导引车（multi-AGV）复杂系统的动态路径规划问题的研究。首先，根据国内外研究现状，对不同AGV复杂系统的冲突类型，通用的规划算法和任务调度方法进行比较和分析。在比较了不同的算法之后，选择了Dijkstra算法作为路径规划算法。其次，建立了总行驶路径中最短路径的数学模型，提出了基于时间窗的多AGV无碰撞路径规划的通用算法。在彻底研究了传统的单车计划和冲突解决算法的缺点之后，建立了规划路径的时间窗改进算法和路径冲突协方差的求解。在VC ++软件上进行的实验表明，改进的算法减少了路径规划的时间，提高了任务的准时交付率。最后，将该算法应用于C企业公司OSIS车间的物料分配。通过比较前后数据的改进，可以确定该方法在实际生产中是可行的，具有一定的应用价值。