Order Picking Optimization And Simulation In "Parts To Picker" Kiva System

The booming development of e-commerce has driven the development of logistics,the requirements for the automaticity and work capacity of distribution center have increased much more.Kiva system comes into being because of traditional automated distribution centers have shown shortcomings for e-commerce order structures such as automatic storage and retrieval system and multi-layer shuttle warehouse.Kiva system has applicated in the Amazon maturely.In recent years,Suning and Jingdong have also begun to use mobile robots to picking.It is very important to improve the efficiency of parts-to-picker picking system.Therefore,this paper studies the optimization strategy of kiva order processing system,which has good practical significance.Firstly,this paper analyzes the order processing mode of Kiva system.It takes several times to move the shelf to the picking station for order processing task.Therefore,decreasing the rack moving frequency to improve the order processing efficiency is our goal,and we think that order processing is mainly influenced by order batching sequence and rack moving sequence.The order batching sequence determined which orders are suitable to be picked in the same picking station,and which order sequence is suitable in the same batch.The rack moving sequence determined the sequence of the rack moving to the picking station.For the order batching problem,the order similarity coefficient is defined to the number of rack required between the two orders,and built the model with the largest similarity coefficient in the system,the improved K-Means clustering algorithm is used to order batching and sequencing.Then,based on the results,the 0-1 integer programming model is established with the least number of rack movements.Finally,the Flexsim software is used to simulate Kiva picking system.On the one hand,the simulation results of the FCFS-based approach and OS-RS optimization are analyzed.The order picking efficiency of OS-RS optimization is improved by 17%,and the number of rack fleet is decreased by 58%.As the scale expands,the scale of the Kiva in the system can be halved.On the other hand,the impact of picking station capacity,storage location assignment and the number of robots based on OS-RS optimization is studied.The results show that picking accuracy and picking efficiency have an effect on picking station capacity;A further reduction of the robot fleet is enabled by the shared storage policy,but this increases the effort during replenishment;if the system have only one picking station,the number of orders plays no appreciable role in terms of required robots,but the robots have the liner relationship with the order structure.