考虑倒箱操作的AutoStore魔方型仓储系统自动引导车双层路径规划  

Automated Guided Vehicle Bi-level Path Planning for AutoStore Cube Storage System Considering Turnover Operation

作　　者：王晓军 杨春霞 陈航慧 范英 白新利 WANG Xiao-jun;YANG Chun-xia;CHEN Hang-hui;FAN Ying;BAI Xin-li(College of Transportation and Logistics, Taiyuan University of Science and Technology, Taiyuan 030024, China;Shanxi Hydra TMMG Mining Tools & Equipment International Co., Ltd., Taiyuan 030024, China)

机构地区：[1]太原科技大学交通与物流学院,太原030024 [2]山西海德拉太矿国际采矿刀具设备有限公司,太原030024

出　　处：《科学技术与工程》2021年第36期15546-15554,共9页Science Technology and Engineering

基　　金：山西省重点研发计划(201903D121176)。

摘　　要：新兴魔方型仓储系统AutoStore中,用于存放货物的料箱垂直堆放在相互拼接但独立的货格内,自动引导车(automated guided vehicle,AGV)除了要搬运目标料箱外,还需负责目标料箱上部阻碍箱的倒箱操作,且活动范围从二维平面扩大到三维空间,均增加了路径规划的复杂性。为了降低路径规划的复杂性,在分解作业流程的基础上,通过建立AGV双层路径规划模型的方法:内层模型以阻碍箱落箱位为决策变量,最小化单个目标箱的倒箱时间;外层模型将内层模型所得倒箱时间为输入参数,优化目标为最小化AGV完成任务时间。通过提出嵌套式启发式算法,将倒箱路径寻优嵌入多AGV多任务路径分配中。研究了不同规模算例,证明了所提方法的有效率和适用性;且仓储规模增大后,目标箱分散度增加,AGV作业时间大幅度增加;AGV数量和作业时间负相关;仓储率提高会增加倒箱概率,增加作业时间;目标箱数目变化与AGV作业时间正相关。结果证明在求解效率方面,主要受目标箱随机生成位置和仓储率影响,但绝对值也仅在10 s左右。In the new magic cube storage system AutoStore,the material boxes used to store goods are stacked vertically in the spliced but independent cargo compartment.In addition to carrying the target material box,automated guided vehicle(AGV)also needs to be responsible for the reverse operation of the blocking boxes on the top of the target material box,and the range of activities is expanded from two-dimensional plane to three-dimensional space,which increases the complexity of path planning.Based on the decomposition of the operation process,a two-level path planning model of AGV was established:the inner model took the blocking position as the decision variable to minimize the dumping time of a single target container.In the outer layer model,the dumping time obtained from the inner layer model was taken as the input parameter,and the optimization objective was to minimize the AGV completion time.A nested heuristic algorithm was proposed,which embeded the optimization of backtracking path into multi AGV multi task path allocation.The efficiency and applicability of the proposed method was proved by different scale examples.Moreover,with the increase of storage scale,the dispersion of target boxes increases and AGV operation time increases significantly.The number of AGV is negatively correlated with the working time.The increase of storage rate increases the probability of dumping and increase the operation time.The number of target boxes is positively correlated with AGV operation time.In terms of solving efficiency,it is mainly affected by the randomly generated position of the target box and the storage rate,but the absolute value is only about 10 seconds.

关 键 词：AutoStore 倒箱 自动引导车(AGV) 路径规划 魔方型仓储系统 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]