A two-stage robust optimization approach for resilient supplier selection and order allocation under disruption risks

The increasing frequency of disruptions in global supply chains poses significant challenges to firms seeking to maintain business continuity and operational stability. Although resilience strategies are widely recognized as essential for mitigating disruption risks, how these strategies should adapt to changing risk environments remains underexplored. To address this gap, this study introduces a supplier procurement framework that integrates normal, flexible, and backup modes, and develops a novel two-stage robust optimization model for resilient supplier selection and order allocation under disruption risks. The proposed model overcomes the limitations of traditional probabilistic methods in modeling disruption risks, and further enhances the representation of disruption risks through an extended formulation that incorporates multiple uncertainty sets. To address the computational complexity of the proposed model, this study develops an improved Benders decomposition algorithm with a multi-cut generation strategy that leverages iteration information to accelerate convergence. Numerical experiments show that the proposed algorithm significantly outperforms both the classical Benders decomposition algorithm and the Column-and-Constraint Generation (C&CG) algorithm in terms of solution quality and computational efficiency. The case study further reveals several practical managerial insights, including the dynamic adjustment paths of resilience strategies under different risk levels, the nonlinear effects of backup strategies, the trade-offs between flexibility and capacity constraints, and the structural patterns of order allocation. This study deepens the understanding of resilience strategy design in dynamic risk environments and offers practical decision support for firms facing disruption risks.