Research On Resource Scheduling And Route Planning Of Intelligent Warehousing System

With the continuous development and improvement of information technology and robot technology,robots have been put into use in many different industries,and warehousing systems have become its important application areas,such as Jingdong’s unmanned warehouse,Ali’s rookie warehouse etc.Robots rely on their excellent perception and recognition capabilities,and based on powerful network transmission technology and efficient information processing technology,laying the foundation for the realization of intelligent warehousing.Taking the intelligent warehouse system as the research background and the intelligent picking system as the research object,the assignment of multi-robot picking tasks,the planning of multi-robot picking paths,and the scheduling of computing resources ware studied in the intelligent picking system.Firstly,the problem of multi-robot picking task assignment of intelligent warehouse system was studied.The allocation mode of multi-robot picking task was analyzed,and the objective function of the shortest picking path and the minimum energy consumption of the robot are established.In the process of solving,this approach of simulated annealing was introduced to address the issue of genetic algorithm,which existed prematurity and lacked of the local search capabilities.This algorithm was improved at the stage of crossover and variation.In order to verify the improved effect of the algorithm,this thesis used three cases,and the effectiveness of algorithm was proved in the problem of multi-robot picking task assignment.Secondly,the multi-robot route planning problem was studied.Based on the reinforcement learning theory,the reinforcement learning Q-Learning algorithm was introduced.To solve the problem of the algorithm,which could not be applied directly in multi-robot system,this thesis redefined the states space and actions space of algorithm.A new reward function is designed to addressed the issue of learning efficiency and conflict in multi-robot system.In order to verify the effectiveness of the algorithm,a robot and multi-robot system were used in different packing environments,and proved the performance,which could give the best route planning scheme.Finally,the computing resource scheduling of intelligent warehousing system was studied.There are frequent calculations from order picking task assignment and the needs for a large number of robot path calculation.In order to quickly complete the calculation in the system,a large number of computing tasks from picking task assignment and robot group path planning ware processed by cloud data center in this thesis.The cloud data center coordinated and scheduled the corresponding computing resources to adaptation with the corresponding computing tasks.In order to achieve reasonable scheduling and ensure the load balance of computing resources,an objective function based on time and cost was established.According to the individual requirements of computing tasks,a two-level scheduling model was established.The two-level scheduling model was analyzed by a numerical example,and its application was verified by a case.