| The splitting and distribution of orders is an extremely important part of the order fulfillment system,especially in self-supporting e-commerce platforms.Orders of self-operated e-commerce platforms often be confronted with a split before being allocated to warehouses.At present,the order splitting method adopted by major e-commerce platforms is based on the principle of “shortest distance”,which causes costs and the number of pickups increasing and inconsistent arrival time of packages,resulting in lower customer satisfaction.In addition,due to the deviation of the front office inventory and the actual inventory,along with the inventory not updated in time,front bunker will short of stock which reveals goods in stock at the beginning.The existing distribution method will delay the processing of the order,which causes the expiration of the customer's time window and punishment.In order to reduce the cost of order fulfillment,offset the deviation of front office inventory from actual inventory and improve customers' satisfaction,the new method of order splitting and distribution is worth studying.Focusing on the particularity of e-commerce orders,this paper conducts research on order splitting and distribution for self-operated e-commerce platforms with“multi-depot” in one metropolitan area.Based on the existing order splitting method—the “shortest distance” principle,a new splitting principle is proposed to improve customer satisfaction and reduce distribution cost,and combine the order splitting process with the subsequent routing path optimization to seek the optimization goal of the minimum order fulfillment cost and ensure customers satisfaction.In the aspect of model solving,the global optimization model is divided into two stages: the first stage is to solve the split and allocation of the order,whose optimal goal is the "minimum quantity of dismantling" and the anamnesis of Set Coverage Problem is used to establish the relationship between order and warehouse.The relationship between each order and each warehouse is obtained.The second stage is to select the distribution route according to the matching relationship,which aiming at the minimum distribution cost and the quantity of vehicle and timewindow is constrained to establish the Nonlinear Integer Programming model,and using the Ant Colony Optimization Algorithm to solve and find the optimal distribution service match between the warehouse and the order.In this paper,the Lingo linear solver is used to solve the small-scale problem.The solution time of the stable case takes about 2 minutes.When the amount of data increases,the solution time will lengthen and the case become unsolvable.In the actual situation,the number of e-commerce orders generated per day is on the number of millions,and the orders is great difference whose demand is varies.Therefore,in the fourth chapter,the improved Ant Colony Optimization Algorithm is used to solve the problem to adapt to the actual scale of the problem.The case part of this paper takes Y company as an example,briefly describes the company's business overview and the existing order delivery problem,and uses the Nonlinear Integer Programming model and the introduced algorithm in forth chapter to optimization the distribution of orders in a certain day,and analyzes the optimal results.Finally,the conclusions and prospects are put forward of this paper. |
