Constructing resilient supply chain for risk-averse buyers by data-driven robust optimization approach

Purchasing plays a crucial role in supply chain (SC) management, directly affecting the production and delivery of enterprises. The serious consequences caused by supply disruptions highlight the significance and necessity of preventing disruption. In addition, the economic panic and anxiety caused by disruptions have prompted SC managers to show a preference for risk avoidance. In this paper, based on diversified procurement and disruption prevention strategies, we study the problem of designing a resilient supply chain network (RSCN) that addresses supply disruption risks and uncertain demand for risk-averse buyers. The excess probability functional (EPF) indicator is innovatively customized to accommodate buyers’ risk preferences. Regarding the uncertainty of demand, we utilize the support vector clustering (SVC) technique based on the given historical data to construct a data-driven uncertainty set and employ it to develop a data-driven robust optimization (DDRO) model. By linearization, epsilon-constraint method, and cone optimization theory, our proposed DDRO model can be reformulated as an equivalent tractable mixed integer linear programming (MILP) model and is solved by a new tailored Benders decomposition (BD) algorithm. Experiments under different settings are conducted on a real-world case, and the obtained results verify the validity of our proposed optimization method.