Research On Multi AGV Scheduling In Intelligent Warehousing System

At present,e-commerce business is at its most prosperous stage and has become an indispensable business model.With the increasing number of online orders,the logistics industry is facing unprecedented challenges,so maintaining logistics efficiency has become critical.In the context of the continuous development of logistics industry and the demand for automation,the application of intelligent warehousing multi-AGV cargo-to-person system has become more and more common.In order to improve the sorting efficiency,how to design multiAGV coordination strategy,so that the multi-AGV system can efficiently complete the order selection,is the current research focus.Aiming at the multi-AGV system,this paper studies order batching,multi-AGV task allocation and path planning from operation management to AGV operation.The main research contents of this paper are as follows:Firstly,considering the order splitting strategy and the task balance of each picking station,the order batching model based on AGV arrival is established with the goal of minimizing the total picking time of AGV travel time and packaging time.A double-layer genetic algorithm(IK-double-layer GA)based on improved clustering is designed to solve the order batch model.The algorithm can solve the problems of low population quality and slow convergence speed of the initial generation of traditional genetic algorithms.The numerical examples show that compared with random batch strategy and non-splitting batch strategy,the number of shelf handling and total picking time are significantly reduced under the order splitting strategy.Compared with the double-layer GA algorithm,IK-doublelayer GA algorithm has faster convergence speed,shorter overall running time,and better overall solution quality and stability performance.Secondly,considering the similarity of the moving shelves required by the picking station and the constraints of the AGV’s remaining power,a multi-AGV task allocation model is constructed to minimize the travel time of the AGV with the longest travel time.An improved genetic simulated annealing algorithm(IGSAA)is designed to solve the model,which can increase the local search ability of GA algorithm.The example shows that compared with the model that does not consider the similarity of moving shelves required by the picking station,the total task completion time of the model proposed in this paper is reduced by 37%,and the number of shelves is reduced by 28%.Compared with the traditional GA algorithm,IGSAA algorithm can obtain better quality solutions in less time and convergence speed is faster.Finally,the map model of intelligent warehouse multi-AGV cargo-to-person system is constructed by using grid method.The multi-AGV collision avoidance strategy based on priority and time window is introduced,and the improved D*algorithm is designed to plan the conflict-free path of multiple AGVs.The example shows that the improved D*algorithm can reduce the number of path conflicts and effectively ensure the reliability of the multi-AGV system.Compared with the improved A*algorithm,the improved D*algorithm can retain a part of the available path,and greatly reduce the AGV traveling distance,diatance deviation and algorithm running time.In summary,considering the similarity of order splitting strategy and moving shelves required by the picking station is of great significance to further improve the efficiency of order picking system,and also provides a reference for intelligent warehousing optimization research.