An Optimization Approach To Order Allocation Problem Of Online Supermarkets With Multi-warehouse In One Metropolitan Area

E-commerce has been leading the fast-developing Internet economy in China, and online supermarkets have become the leader of the fast-developing Chinese e-commerce. As a new type of electronic business model, online supermarkets usually sell over millions kinds of goods, which brings great convenience to the customers and great challenges to its operation development. With the features of large number of multi-item orders and multi-warehouse in a metropolitan area, online supermarkets need to allocate over ten thousands of orders to warehouses every day. Due to the storage of different warehouses, many orders are split and packaged into multiple packages. This unreasonable allocation of orders often leads to split delivery, which not only causes the high logistics cost, but also results in the customers' dissatisfaction and complaint. Therefore, the problem of order allocation becomes the key problem that needs to be solved urgently for the survival and development of online supermarkets.Order allocation is first core stage of order fulfillment in online supermarkets, which will directly influence the following stages in the whole order fulfillment process. The order allocation of online supermarkets with multi-warehouse in one metropolitan area, not only has to consider order splitting and allocation in multi-warehouse, but also considers the delivery problem. Therefore, the order allocation problem is a complex optimization problem that needs a combination of qualitative and quantitative analysis.In terms of this problem, the main work of this research is as follows:(1) Complexity analysis of the order allocation problem. This paper analyses the key factors of orders, warehouses, distribution stations and vehicles in detail when considering the order allocation problem. Then the complexity of this problem is analyzed in the modeling, solving and other aspects.(2) An optimization model of order allocation. With the objective of minimizing the cost of logistics, considering the constraints of multi-item order allocation, delivery and other decision variables, this paper presents the nonlinear mixed integer programming model.(3) A heuristic algorithm for the order allocation problem. Considering the characteristics of this problem, this paper proposes a heuristic algorithm by introducing the state-space search theory in artificial intelligence and the local search algorithm. The heuristic called "split-consolidate-optimize" is based on the idea of reducing the solution space, which combined the improved breadth first search algorithm, the improved C-W saving algorithm and the improved the local search algorithm.(4) Application research. This paper verifies efficiency of the proposed model and algorithm in solving order allocation problem, especially for large-scale order allocation problem, in the online supermarkets with multi-warehouse in one metropolitan area based the generated numerical cases with the characteristics in reality.This paper provides a new method and tool for multi-item order allocation in online supermarkets with multi-warehouse in one metropolitan area, which is helpful to improve the technicality and practicability in order fulfillment operations of online supermarkets. In practice, the theoretical results of this paper can be used to guide the actual operation of dispatchers in B2C business enterprise, which is helpful to reduce the logistics costs of orders and improve the efficiency of enterprises' order fulfillment operations.