Research On Demand Forecasting And Replenishment Based On Modular Order Picking

With the development of Internet, more and more people are used to shopping online. E-commerce companies receive a lot of orders everyday, and these orders are usually small and diversified. This phenomena directly leads to the increase of sorting complexity and time, and further affects inventory turnover and the assurance of timely delivery. In order to meet the diversified demand of orders quickly, this thesis presents a thought of modular order picking.Namely analyze orders to find out the commodity combinations which repeat in a large number of different orders, then choose effective combinations and define them as modules, and the module is a new sorting unit, so we need to assign storage location and design order picking policy for them. At the same time, we need to package modules simply for convenient storage and picking, so the problem of module combination and replenishment is involved. The Above content can be divided into three subproblems: the determination of effective modules, the adjustment of sorting strategy and the replenishment planning of modules.This thesis is focused on the third problem.Firstly, a large number of relevant literatures are studied for the subsequent research. Then, the problem was devided into two sub-problems: module demand forecasting and the determination of module replenishment lot-sizing. Firstly forecast modules demand, and then use the results to determine the replenishment batch of modules.With these two sub-problems, this paper aims to minimize module replenishment cost, at the same time, ensure smooth sorting and satisfy capacity constraints of module replenishment staff and storage capacity limits. For the first sub-problem, a suitable module demand forecasting method was putted forward based on the reference of some existing forecasting methods, and an instance analyse was conducted to verify the method based on the data of a real enterprise. For the second sub-problem, on the basis of some traditional lot-sizing problem models and enterprise actual operation situation, a model of module combination lot-sizing under capacity constraints is putted forward.