Research On Task Allocation And Path Planning Of Robotic Mobile Fulfillment System

In the era of "Industry 4.0",robot technology has become an important methods to assist the transformation and upgrading of LaborIntensive industries.In recent years,the Ministry of Industry and Information Technology has pointed out that it is necessary to build intelligent logistics systems with a focus on robots,and improve the digital level of commercial logistics.Robotic Mobile Fulfillment System(RMFS)is a new type of picking solution emerging under the tide of intelligent transformation of traditional logistics facilities.It replaces the traditional operation mode of "human picking" with "robot picking",greatly improving sorting efficiency and accuracy,while significantly reducing labor costs.However,scheduling large-scale robots to collaborate safely and efficiently is still a hot and difficult topic in current industrial and academic research.This paper takes RMFS as the research object,studies the task allocation and path planning methods for robots under a given set of tasks,and independently builds an RMFS simulation system to validate the proposed algorithm,providing theoretical guidance for improving the sorting efficiency of RMFS.The main work of the paper is as follows:(1)Aiming at the dynamic task time-consuming problem and the limited capacity of the seeding wall in RMFS,a task allocation method based on a hybrid heuristic algorithm was proposed.A task allocation model is established to minimize the maximum makespan under order priority constraints.A maximum makespan generation scheme is designed considering robot acceleration and deceleration,turning,lifting and lowering of shelves,queuing area waiting,and seeding wall capacity constraints.A hybrid heuristic algorithm MEPCALNS is proposed,which introduces a search population based on catastrophe operators and an elite population based on probability update based on ALNS.The search efficiency and optimization effect of ALNS are improved,and numerical experiments demonstrate the effectiveness and superiority of MEPCALNS.(2)Aiming at the problem of insufficient estimation of robot kinematics characteristics in RMFS when planning robot clusters’paths,a multi-agent path planning method based on multi-agent offline planning and multi-agent online coordination was proposed.In the offline planning stage,based on the classic MAPF solver RRA*-WHCA*,an improved algorithm KRRA*-WHCA*is proposed.By introducing an auxiliary coordinate and a standard count comparison table,the minimum turning time is accurately solved,optimizing the evaluation of estimating costs,and improving the routing effect of RRA*-WHCA*.In the online coordination stage,a coordination algorithm with kinematic constraints is proposed based on the discrete timestamp coordination algorithm.Through the resource application and release mechanism,the problem of decisionmaking timing under continuous speed constraints is solved;Through the deadlock detection and resolution mechanism,it ensures that the robot cluster does not have local circular waiting.(3)In order to verify the effectiveness of the proposed algorithm in a dynamic simulation environment,a dedicated visual lightweight RMFS simulation system was independently built.The system adopts Agent based discrete event system simulation technology,uses the.NET6 C#and WPF,and implements functional modules such as scenario configuration,task allocation,path planning,and simulation run.The proposed task allocation algorithm and path planning algorithm are validated by simulation experiments using the RMFS simulation system.The simulation results prove the advantages of MEPCALNS and KRRA*-WHCA*compared to their respective classic algorithms.