6. Conclusion and suggestions
Lower costs, commercial treaties, new markets, developing communication and internet opportunities are some of the advantages of globalization. However, cultural diversities, standardization difficulties, political instabilities and extended distances make SCs more vulnerable to risks. Today, SCRM ensures not only cost advantage but also competitive advantage. As presented in the literature review, the number of quantitative and model-based studies is more limited when compared to the number of qualitative and empirical ones. Within this context, a proactive planning procedure is proposed in this paper. The aim of this procedure is to take necessary precautions against the risky suppliers and decrease the level of damage in case a disruption occurs. The first stage of the proposed procedure is obtaining the initial procurement plan via linear programming with the objective of cost minimization, and the second stage is revising the initial procurement plan by integrating the risk criteria into the planning process. Risk assessment is conducted so as to identify risk profiles of all suppliers in the first step of this second stage. In the second step, the product quantities in the initial procurement plan are proportioned to risk profiles of the suppliers, and then the quan tities to be kept in a supplier and those to be transferred to less risky suppliers are identified. In the third step, a product transfer Table 19 Probability - impact matrix. linear programming model is established and in the last step, the new procurement plan is created by solving this linear programming model. This procedure is unique in that risk is quantified and included in the model not in terms of cost but as a profile value and it proposes a transfer of product strategy. This transfer plan is made before the order and suppliers receive the final product order prepared according to the cost and risk criteria. In the numerical example, the model is tested 10 times with a randomly generated hypothetical data set and verified. The cost increase incurred by the inclusion of risk criterion into the model is analyzed and a 95% confidence interval for the population mean of cost increase is observed to be between 1.32% and 3.88%. This cost increase is likely to be less than the cost that will be encountered in the case that a disruption occurs.