Operational Research Of Supply Chains Based On Demand Characteristics

It is the most important object for a supply chain to satisfy the market demand. As the driver of supply chain performance, demand affects the production, inventory management and sale activities of each firm in supply chains. Under different circumstances of supply chains, demand will demonstrate different characteristics and managers should correspondingly adapt different management methods. The goal of this doctorial dissertation is to analyze the way that different characteristics of demand affect supply chain performance and then develop policies to improve the efficiency of supply chains.Based on uncertainty, information asymmetry, ambiguity and competition of demand, this dissertation analyzes the performance of supply chains and the approaches to improve it. Specifically, the content can be abstracted as follows.(1)When the demand is uncertain, we analyze the propagation process of demand information. The propagation process of demand information can be viewed as a transformation process of demand information by inventory management policies of each firm. We classify inventory management policies into linear inventory management policies (LIMP) and non-linear inventory management policies (NIMP). The feasible condition of LIMP is presented and propagation rule of demand information under LIMP is developed. And the relationship of LIMP and the Bullwhip Effect is analyzed and the optimal LIMP to eliminate the Bullwhip Effect is designed.(2)To respond the uncertainty of demand, we develop an optional contract to coordinate the manufacturer and the retailer. A three-stage dynamic optimization model is developed to describe the option contract. The retailer's ordering process is separated into three stages with demand information reveal. We derive the retailer's optimal ordering policy and prove the validity of the option contract. And the effect of life cycle of the product, capacity of the manufacturer and the market information on the retailer's optimal policy is discussed.(3)When the demand information is asymmetric, we analyze the impact of information asymmetry on supply chain performance and then present a sequential belief revision method to eliminate information asymmetry in supply chains. First, by comparison of four types of games, the impacts of information asymmetry and information uncertainty are identified. Second, we assume that the retailer has private information about demand and the supplier only partially knows the retailer's optimal decision. The supply chain performance is modeled as a multi-stage Bayesian game with observable results. We show that with the sequential belief revision method, the supplier's belief will converge to the retailer's private information and the game equilibrium is robust.(4)When the demand is unknown, we analyze the management approaches for a single retailer and the whole supply chain. First, the inventory management problem of a single retailer is analyzed. We prove that the traditional approach of replacing parameters by their estimates will make the newsvendor away from the optimal profit level, leading to a suboptimal result. And then two revised application methods are provided. Second, the Bayesian inventory method (BIM) is analyzed in the circumstance of supply chains. We show that BIM may lead to pooled information effect. The definition of pooled information effect is presented and sufficient conditions for this effect are developed. And the source and countermeasure of pooled information effect are discussed.(5)When the demand is competitive in multi-channel supply chain, we analyze the optimal channel selection policy for a capacitated manufacturer. Assume that the manufacturer has limited capacity and she may distribute her product through a wholly-owned direct channel and an independent retail channel. Demands of the channels are substitutable, which leads to competition between them. A three-stage dynamic game is modeled to describe the supply chain performance and the game equilibrium is derived. We prove that when the capacity is very tight, the direct channel does take priority over the retail channel; otherwise, channel selection decision depends on the available capacity and the substitution rate. We show that under certain conditions, the manufacturer will eliminate the retail channel, even hoarding her capacity when the capacity is still ample.To make supply chains coordinated and satisfy the market demand, managers of supply chains should grasp the characteristics of demand. Based on four characteristics of demand, this dissertation studies four types of management methods, which will shed light on practical supply chain management.