Research On Multi AGV Real-time Dynamic Scheduling And Three-dimensional Packing In Intelligent Picking System

With the rapid development of e-commerce and express industry,the demand for packaging products has become increasingly large and diversified.Under the trend of mass customization,intelligent logistics equipment is actively applied in the industry to realize rapid response and efficient and accurate picking of orders.In this paper,AGV scheduling based in the plant logistics and three-dimensional packing problems are studied,to achieve efficient material handling,packing operations and the effective connection of the two,reduce the cost of logistics transportation.The main work and research contents of this paper are as follows:This study proposes a method of task allocation based on adjacency combination and the shortest path principle.Three priority rules for the mobility of AGVs between nodes are designed,and a transportation strategy that combines single and two-way paths is proposed to reduce computational complexity.By combining with Dijkstra’s method,the authors develop a method to prevent deadlocks and collisions between multiple AGVs based on a timetable of reservations that hierarchically handles conflicts among nodes in multiple stages.Such constraints as AGV congestion or deadlock weaken the effectiveness of the shortest distance rule-based solution.Based on the above,a heuristic search method based on variable neighborhood search is further proposed to optimize the problem of multi-AGV task assignment,and a corresponding theorem is given to avoid the generation of unfeasible solutions by the neighborhood operators and improve the efficiency of the solution.The results of experiments show that the proposed method can adequately solve the problem of scheduling multiple AGVs in a large and dense network.Then,a class of heterogeneous three-dimensional packing problem strictly following the first in and last out constraint is studied.According to the actual scheduling results of large-scale AGVs,a single container loading algorithm based on greedy heuristic multi-layer tree search is proposed.The algorithm adopts the idea of block loading,greedy heuristic searches each layer,uses the most suitable block for loading,and obtains a loading sequence after unable to continue loading.Combined with the single container loading algorithm,the container with the highest loading rate among all specifications and the corresponding loading sequence are successively selected to generate the initial solution of multi container loading.Based on the above algorithms,a large neighborhood heuristic search algorithm including five removal operators and two repair operators is further designed to optimize the multi container loading problem,which can effectively improve the quality of the packing solution.Some examples are given to demonstrate the effectiveness of the proposed algorithm.Finally,based on the above research results,a three-dimensional packing system is developed based on C#language,MATLAB and MySQL database to dynamically display the three-dimensional packing process and update the loading steps step by step,to provide guidance and basis for on-site loading operation.