Analysis of a supply chain with forecast evolution and supply disruption

Abstract

In this thesis, we analyze two supply chain models. In the first model, we develop a centralized supply chain consisting one warehouse and two retailers. The aim of this model is to find the total optimal order quantity for the warehouse. We assume that the warehouse faces yield uncertainty. At the beginning of the period, both retailers place orders and the warehouse starts supplying them. However, if there is not enough on hand inventory at the warehouse at the time of order, then it should allocate it to the retailers. After receiving units by the retailers, stochastic customer demand is observed and inventory costs are occurred. We develop and present a formula to calculate the total optimal order quantity that minimizes the total system cost and the formula can be solved in numerically; therefore we examine results in the numerical analysis section. In the second model, we consider a decentralized supply chain with one manufacturer and one retailer under a wholesale price contract. The goals of this model are to find the total order quantity for the retailer, the total advance production amount for manufacturer and the wholesale price. The manufacturer is subject to yield uncertainty. In addition, retailer can update its forecast after obtaining better information. At the beginning of the period, both of them agree on wholesale price and the manufacturer starts producing before receiving the retailer’s order. When the retailer obtains forecast update, it gives total order quantity to the manufacturer. If advance production quantity is not sufficient to meet the retailer’s total amount, then the manufacturer makes outsourcing for the remaining quantity. After receiving total units by the retailer, customer demand is satisfied with a retailer price and the market uncertainty is realized. We develop and present an explicit formula for the retailer’s problem. The manufacturer’s problem can only be solved in numerically and accordingly the wholesale price can. We also examine a numerical analysis for the model.