Research On Scheduling Strategy And Path Planning Based On Warehouse AGV

Automatic guided vehicle(AGV)acts very important role in current logistics warehouse.It is widely used in the warehousing and logistics industry because of its small body,high flexibility and strong operability.In view of the logistics warehouse with more packages,less AGV and complex road conditions,this paper aims to improve the efficiency of the system,reduce the overall working time as the guidance,uses the simulation experiment to simulate the real environment,uses the improved task allocation algorithm and path planning algorithm,solves the related problems,and achieves certain innovation.The research contents of this thesis has several parts below:(1)In actual production,multiple packages usually need to be processed by an AGV,resulting in the problem of task allocation sequence among multiple packages In this paper,the original opt algorithm is easy to fall into local optimization and takes a long time to solve the problem of task allocation.A k-opt optimization algorithm based on data structure LRU cache is proposed,To speed up the access to the recently accessed data,K-opt adds multi pair optimization algorithm to avoid the algorithm uncertainty caused by falling into the local optimal solution and increase the possibility of finding the relative optimal solution.(2)In order to get closer to the actual warehouse working environment,this paper uses the grid map to simulate the real warehouse environment.On the basis of retaining fixed obstacles,random obstacles are added.At the same time,the AGV size and obstacle size are considered in the judgment of obstacles,which also provides a better planning path.(3)For the problem of path planning,this paper proposes two solutions to the problems of a*algorithm,such as long time of path finding and many inflection points of path generation.By using the improved evaluation function to calculate the proportion,the experiment shows that the time of route finding is significantly shortened.Smooth the generation path,select the key generation points for global optimization,local optimization for each key point,and finally connect all the key points into the final generation path.The simulation results show that the inflection points in the generation path are significantly reduced,and the curve fitting is more smooth.