工厂环境多AGV动态调度系统的组合优化  

Multi-AGV Dynamic Scheduling System and Combination Optimization Algorithm for Factory Environment

作　　者：李睿 朱笑笑[1] 栾楠[1] LI Rui;ZHU Xiaoxiao;LUAN Nan(Shanghai Jiao Tong University,Robotics Institute,Shanghai 200240,China)

机构地区：[1]上海交通大学机器人研究所,上海200240

出　　处：《机械设计与研究》2019年第6期37-42,共6页Machine Design And Research

摘　　要：为解决某玻璃厂对多自动导引小车(AGV)的需求,提出了一套工厂AGV动态调度系统的软件架构。以具体的玻璃运输环境为研究对象,考虑路径冲突,AGV停车等一系列随时可能出现的现实因素,引入时间窗队列解决路径冲突问题,并利用时间窗优化Dijkstra算法。通过预处理地图信息减少算法计算时间复杂度。基于AGV系统的分布式特性提出多AGV报价竞争的路径敵合算法,通过计算特征量将相似路径的订单融合成一条优化路径,缩短订单池完成的总时间。In order to solve the demand of a glass factory for multiple automatic guided vehicles(AGV),a software architecture of a factory AGV dynamic dispatching system was proposed.Taking the specific glass transportation environment as the research object,considering the path conflict,AGV parking and other real-time factors that may appear at any time,the time window queue is introduced to solve the path conflict problem,and the time window is used to optimize the Dijkstra algorithm.The time complexity is calculated by preprocessing the map information reduction algorithm.Based on the distributed characteristics of the AGV system,a path fusion algorithm for multi-AGV quotation competition is proposed.By calculating the feature quantity,the orders of similar paths are merged into one optimized path,and the total time of the order pool is shortened.

关 键 词：多自动导引小车 动态调度系统 时间窗 路径融合 DIJKSTRA 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]