Risk-averse supplier selection and order allocation in the centralized supply chains under disruption risks

Abstract

This paper proposes a mixed-integer non-linear programming (MINLP) model for the integrated supplier selection and order allocation in a centralized supply chain considering the disruption risks and a risk-averse decision-maker. In order to capture a realistic scenario of considering the geographical characteristics of the suppliers, we assume that the suppliers belong to two regions: the buyer’s region (domestic suppliers) and outside of the buyer’s region (foreign suppliers). Considering this realistic feature, the supply chain might face two types of disruption risk: first, local disruption risks which might uniquely occur inside each supplier such as equipment breakdowns, and second, regional disruption risks that might occur in the region of the suppliers located in the same geographical region such as natural hazards. We formulate the problem considering a risk-neutral decision-maker as a benchmark, and then a risk-averse model is presented. In the latter case, we apply two types of risk assessment tools introduced in the finance literature to analyze the decision maker’s behavior: value-at-risk (VaR) and conditional value-at-risk (CVaR). We show that developed models are non-convex programming, and therefore, we apply the particle swarm optimization (PSO) algorithm as the solution approach. We also compare the developed PSO algorithm with the Genetic algorithm (GA) and the commercial GAMS solver to verify the efficiency of the solution method. The computational experiments indicate the impact of the decision maker’s attitude on the supplier selection and the order quantity.