Multi-product Dual Sourcing Procurement Strategy Research Based On Demand Information Updates

With the development of science and technology and the increasing complexity of the market environment, the market demand has shown increasing uncertainty. And uncertainties in supply and demand often result in massive risks for a business. To reduce the gaps between supply and demand and to hedge supply chain risks, multiple-source sourcing is often adopted. Multi-source sourcing means source from different channels or markets. In this paper, we study the multi-product dual sourcing based on demand information updating. Demand forecast information updating is a method which updates the prior demand forecast by using the realized demand in the first stage so as to obtain the posterior demand forecast. Retailer has two opportunities to order multiple products from two markets—contract market and spot market. The first order comes from contract market which has a long lead time, a lower wholesale price, while the second order comes from spot market, which has an instant lead time and a higher price. In one case, the two orders are made before the beginning of the sales season, thus the retailer can use the market signal observed between the two orders to update the demand forecast and determine the order quantities of each products separately from the two markets in the balance of the budget and demand uncertainty. Another case is the first order is made before the beginning of the sales season, the second order is made in the middle of the sales season and then the season is divided into two stages. Retailer can update the demand forecast based on the previous sales situation, and then modify the order quantities of the next phase, so as to obtain the maximum benefits. Here we consider the situation of demand information censored (DIC). Demand information is censored when sale quantities of stock are less than the market demand and then customers’demand can’t be met, which may lead the real market demand not fully acquired, also called demand information partially observed or incompletely observed referred to some articles. Demand information censored is not only a common phenomenon in enterprise management, but also an important feature of modern business environment. In practice, retailers tend to make the sales data equal to the demand, but when the demand is greater than the supply quantities, the sales data is not the actual demand. When retailers make procurement decisions without consideration of the situation of DIC, then the demand data obtained before is not accurate, and if retailers use these data to update the demand, it may be a vicious cycle and result in a greater loss. So it is necessary to consider the possibility of DIC when using the information updating to update demand forecast.In this paper, we established models according to the situations of two orders occurred before the start of the selling season or the second order occurred in the sales season and whether consider the situation of DIC. Then we carried on the solution calculation and analysis, in addition, we also set up optimization algorithm and gave the example analysis.