Research On Order Batch Optimization Method Based On Cluster Analysis

With the rapid development of e-commerce,the sales model of traditional retail is gradually transformed from B2 C to C2 M.The size of the orders becomes small,but the number is huge.What puts forward higher requirements for the operational capacity of warehousing logistics.In the warehousing enterprises with manual work,the sorting time accounts for 45% of the whole operation time.The ratio indicates that as the key process of warehousing logistics,sorting efficiency has an important impact on the operation and service level of an e-commerce enterprise.Therefore,it is particularly important to study the storage assignment and the order batching to improve the efficiency of the picking operation.Warehousing centralization is a common inventory management mode to improve warehousing logistics.Under the condition,the research and analysis of effective zoning strategy make the warehousing centralization more effective.This paper chooses the parallel zoning operation mode,establishes the allocation model of goods location with the objective of minimizing the rate of tardy job,calculates the degree of association of goods by using association rules,and realizes the optimal design scheme of storage assignment.The order batching strategy is studied under parallel partition sorting.Aiming at the actual problem of long order processing time caused by the huge difference of picking completion time in different zones,a mathematical model of order allocation based on quantitative order was constructed.The model takes processing time as objective function,and taking order segmentation and equipment resources as constraints.Aiming at the problem of workload balance in batch model,DBSCAN algorithm and K-Means algorithm are combined to solve the batch model,and the optimal batch result is obtained.The result is substituted into the mathematical model of order allocation and the optimized order set with the shortest time.The model of order allocation and batching is validated based on real environment and case data.The effects of parallel partition optimal allocation strategy and random strategy on order batching are compared and analyzed.The results of density-based KMeans clustering and traditional K-Means algorithm on order batching are evaluated and analyzed.The experimental results show that the parallel partition sorting system is effective.In the case of large data sets,K-Means clustering algorithm based on density can make full use of picking equipment and personnel to reduce the number of batches,shorten the order picking completion time,and effectively shorten the picking distance and time on the basis of comprehensive utilization of storage location optimization strategy,improve the operational efficiency of warehousing logistics.