Research On The Picking Route Planning Of Mobile Racks Based On AGV

Advanced information technologies such as Artificial Intelligence,Internet of Things and Cloud Computing have promoted the transformation and upgrading of the traditional logistics industry to the modern logistics industry,especially the development of intelligent robots,which has greatly improved the efficiency of all links of logistics and effectively reduced the total logistics cost.In recent years,a new model of "Parts to picker" picking operations has begun to be adopted inside distribution centers.In this mode,an AGV(Automatic Guided Vehicle)is used to transport the racks to the workbenches,and the pickers will pick out the goods.After visiting several workbenches,the racks will return to the original position,and then the AGV will move to the next rack until all orders are completed.For this application scenario,this paper studies the problem of AGV picking route planning.The main research work is as follows:(1)On the basis of summarizing the domestic and foreign research status of vehicle routing problem with pickup and delivery,AGV scheduling problem and picking route planning problem,the types of picking operation,"Parts to picker" picking operation mode,storage location allocation strategy,storage layoutand AGV environment modeling method are systematically analyzed.In this paper,an AGV-based mobile rack picking route planning problem(MMPRP)was proposed,which was characterized by the AGV selecting racks and accessing multiple workbenches.The relevant assumptions,constraints and topological map of environment modeling were given,and the traveling distance model with the objective function of minimizing the total traveling distance of AGV was established.(2)According to the characteristics of MMPRP problem,a hybrid ant colony algorithm(MACO)is designed to solve it.The algorithm on the basis of the basic ant colony algorithm,and improved the transition rule and pheromone update methods,combination of factors such as the distance from the rack to the workbench,the type of goods on the rack,etc,designed a rack selection rule suitable for the MMPRP problem.Max-min ant colony strategy,elite ant strategy and path saving factor were introduced into the pheromone updating rules.In addition,six local optimization operators were designed to improve the performance of the algorithm,including merging single access nodes,removing terminal AGV access requirements,replacing racks with low satisfaction,realigning the access path,removing nodes of the longest segment,and exchanging two rack segments.(3)Forty-eight instances were built to verify the validity and applicability of the MACO algorithm,and the effects of random storage strategy,turn-based storage strategy,and the layout of the workbenches on the solution were explored.Experimental results show that the proposed MACO algorithm,which integrates six local optimization operators,can effectively solve MMPRP problems,and has better ability to explore the optimal solution and search the solution space than the ant colony algorithm alone.Among the instances of about 60 kinds of commodity sizes,the MACO algorithm found the optimal solution.Compared with the ACO algorithm alone,the average improvement rate in the moving distance model was as high as 20.43%.