Research On Optimization Of Storage Location Assignment Strategy Considering Product Association Structure Of Consumers' Market Basket

With the increasing competition among e-commerce platforms,how to improve shopping experience,especially to solve the problem of satisfaction delay caused by slow logistics,has become the key to improve the operation efficiency and competitiveness of e-commerce platforms.Due to the massive data generated by consumers who shop online,the practice of e-commerce warehouse management has begun to pay attention to mining market basket data to achieve accurate replenishment.Meanwhile,the optimization of storage location assignment to improve the efficiency of order picking has also aroused the interest of researchers.Based on association rules mining,the method of storage location assignment such as Association Seed Based Heuristic(ASBH)puts forward the idea that items frequently appear together in the same order should be placed closer in the warehouse.Simulation experiments show that this method has a significant improvement in order picking efficiency compared with methods which don't take product association as consideration.However,association rules based method only considers the relationship between any two products in market basket and doesn't make full use of the complex relationship structure formed by co-purchase pattens.This research proposed a product association structure based(PASB)method which considers the product association structure of consumers' market basket.PASB method integrates data mining,network analysis and optimization technique to solve the storage location assignment problem.Aiming at the optimization of order picking efficiency,PASB takes the structure characteristics of co-purchased pattens as consideration to perform a better storage location assignment result.The steps of PASB are as follows:(1)Performing association rules mining algorithm to identify association pattens by using historical data of market basket from the sales end,using association rules to build the product association network.(2)Using Louvain algorithm to identify the subgraph structure in the product association network and obtain product groups in which products are strongly associated with each other.(3)Building the 0-1 programming model of storage location assignment problem which is to maximize the association index.The product groups information is integrated into the programming model,and the final result of storage location assignment is obtained by solving the model through the optimizer.The simulation experiment of PASB method takes the improvement degree of travel distance for picking as the evaluation criterion and compares the effects of PASB method with ASBH method,random assignment method(Rand)and the method that only consider product purchase frequency(Freq).The experiment results show that:(1)Compared with Rand method and Freq method,PASB has obvious advantages.(2)Compared with ASBH method,the effect of PASB is affected by the warehouse space constraint.We proposed Disruption as an indicator of PASB suitability and proved that PASB is better when Disruption is higher.The simulation results show that PASB provides a new way to solve the problem of storage location assignment by considering product association,which has certain promotion value and practical significance.