Multiple perspective models of manufacturer-supplier risk

Supply networks are becoming more coupled, inter-dependent, complex, and dynamic because of business trends and global competition. In a competitive environment, suppliers have large and direct impact on cost, quality, technology, and time to market of new products. In this dissertation, we study this complex, dynamic, risky supply chain environment from both an Original Equipment Manufacturer (OEM) and a supplier perspective. We first explore global sourcing risks from an OEM's point of view and develop models to analyze and manage risk. We then study the impact of long-term relations between a supplier and OEM on bidding decisions and develop sequential bidding models. Efficient management of global sourcing risk creates a competitive advantage and reduces the probability and/or impact of detrimental events. In this dissertation, we develop an emerging market sourcing risk analysis and management model for a global automotive company to support its global sourcing decisions. We use simulation to analyze risk and evaluate risk mitigation plans. The model not only helps decision makers assess risk, it also guides them in developing risk mitigation plans. We define a number of sourcing risk management strategies and then use optimization to find an optimal portfolio of the risk management plans by reflecting the decision maker's risk attitude. The model has been verified and validated through the real world case studies. We also develop models of sequential bidding on product design and manufacture for a tier one supplier. The initial contract and its successful implementation impact the organizational relationship between the supplier and the buyer which directly affects pricing, profit margins, and the probability of winning a subsequent contract. We examine the sequential bidding problems for finite and infinite time horizons. We use backward dynamic programming to derive optimum bid prices for finite time horizon models. Infinite time horizon models relaxed some of the assumptions we made to derive analytical solution for finite time horizon bidding models. We use Markov Decision Process to derive the optimum bidding strategy for each state.