Research On Optimization Problems Of AutoStore System

With the rapid development of the mobile Internet,the development of the online retail industry has become more and more prosperous,and the efficiency of order picking has received much attention.At present,the AutoStore system,a"Goods-to-person" demolition and picking solution that can significantly improve labor efficiency,order accuracy and order utilization,is emerging.The system was discovered only in the last ten years,and many optimization studies have not been completed.Research on optimization problems of AutoStore system can improve the responsiveness and operating efficiency of the logistics system.It has important theoretical and practical significance.Based on this,this paper studies the optimization of key operating strategies in the AutoStore system.The specific research contents and results are as follows:This paper identified six operating modes of the AutoStore system,and built a system performance indicator evaluation model based on semi-open queuing network for each mode.According to the actual requirements of the AutoStore system,this paper considers the robot rules for special robots and general robots,dedicated storage and shared storage policy,and also considers immediate and delayed reshuffling policies in shared storage.Combine different strategies to determine six models for research.First analyze single-line orders and then expand to multi-line order mode.The queuing network is solved to analyze the performance of the AutoStore system.First,the approximate mean value method is used to solve the service node time in the queuing network.Then the matrix geometry method is used to approximate the network to calculate the system performance parameters.Finally,a simulation model was established to verify the theoretical model.The research results show that:when the number of robots is small,the general robot rules are more conducive to improving the system throughput performance,while when the number is large,the special robot rules are more suitable.Although the specified storage strategy is more conducive to improving the order picking efficiency of the system,it requires more storage space and higher investment costs.And the system can achieve better throughput performance by using immediate reshuffling rather than delayed reshuffling.The experimental design was used to expand the structure design and resource allocation optimization in the AutoStore system.The results show that:in terms of structural design,high grids are more suitable for dedicated storage policy,while flat grids are more suitable for shared storage policy.In addition,the optimal length-to-width ratio r*of the grid shelf under the general robots rule is about 0.65,and it is slightly less than r*about 0.52 under the special robots rule.In terms of system resource allocation,under the rules of special robots,the average order throughput time decreases as ? increases.Under the general robots rule,the throughput time decreases first and then increases.And the ratio of the existence of robots and picking stations ?*is about 12,when it is less than ?*,the system performance of general robot rules is better than that of special robot rules,and vice versa when it is larger than ?*.