Multi-Objective Decision Support System for Global Supplier Selection

In any large organization, millions of dollars are spent on purchasing. Raw material cost accounts for 40-60% of production costs for most US manufacturers. In fact, for the automotive industry, the cost of components and parts from outside suppliers may exceed 50% of sales. It is vital to the competitiveness of most firms to be able to keep the purchasing cost to a minimum. In todays competitive operating environment it is impossible to successfully produce low-cost, high-quality products without good suppliers. An organization cannot afford a strategy that treats all items, products and services in the same manner. More and more firms are finding that the only way to align the procurement function with the firms overall goals is to segregate purchasing into different segments based on supply strategies, supply tactics and supply management approaches. This process is known as supply segmentation. Supply segmentation technique provides a mechanism for distinguishing among different items and services that are purchased by a firm with the goal of developing specific strategies to meet the needs of the organization.In this dissertation we segregate the purchasing process into a 2x2 matrix (4 quadrants) with the cost of products on x-axis (Low/High) and supply risk on y-axis (Low/High). The objective is to develop supplier selection methods for each quadrant of the matrix. For each quadrant, we formulate a different optimization model and solve it using multi-objective optimization methods.For the Tactical products (quadrant 1, low cost, low risk) we consider a multi-period planning horizon problem. The buyer forecasts the demand of each product for each time period. The objective is to decide what products to buy from which supplier and in which time period; we also incorporate product bundling as a form of quantity discount.Leverage items (quadrant 2) have a high price but low supply risk. Hence, the primary objective for supplier selection in this category is to reduce the price of the purchased products. We model and solve the problem as a multi-objective problem with price, lead-time and quality as three conflicting criteria. Typically for leverage items there are many suppliers to choose from; hence, we first pre-qualify the suppliers using cluster analysis. Pre-qualification is the process of reducing large set of initial suppliers to a smaller set of manageable suppliers.Critical items represent quadrant 3 (low cost, high risk). Although critical items represent a small portion of the total cost, they present a very high level of service disruption. In this quadrant we suggest the organizations seek supplier base globally. Hence, along with cost we incorporate risk factors in the supplier selection process. We assign a global supplier, primary domestic supplier and primary secondary supplier for every product with the aim of minimizing both cost and risk.Quadrant 4 represents strategic items (high cost, high risk) that require long-term partnership with the suppliers. Multi-criteria decision making methods with multiple decision makers are most suitable for this quadrant. We use Analytic Hierarchical Process (AHP) and Total Cost of ownership (TCO) to measure qualitative and cost criteria. A pairwise comparison method is used to combine AHP and TCO objectives to determine the best suppliers.Finally we introduce the use of Radio Frequency Identification (RFID) in Supply Chain Management and present the benefits of implementing RFID through a real-life case study.