A resilient robust risk-averse disaster relief supply chain network

Natural disasters cause significant human and economic losses worldwide, highlighting the need for effective preparedness and integrated disaster relief supply chain (DRSC) networks. Therefore, this study designs an efficient DRSC network within a multi-period pre-disaster planning time horizon (PTH), operating within an uncertain environment while considering the perishability of relief items (RIs) and the gradual allocation of a limited budget prior to a disaster. To this end, a novel risk-averse two-stage stochastic-robust programming model is proposed. In each planning period, the first-stage decisions involve strategic and tactical pre-disaster activities, including warehouse location and inventory management, while the second stage focuses on post-disaster operational decisions related to the re-procurement and distribution of RIs. To facilitate the procurement of RIs, two innovative contractual agreements are developed, which are integrated with multi-sourcing and enable two-part buybacks, the selection of multiple reservation options, and quantity-based discounts on specified conditions. By introducing a novel service utility metric, the model seeks to optimise fulfilled demands and deprivation costs. Furthermore, a light robust optimisation approach is employed to solve the model efficiently. The model’s effectiveness, efficacy, and practical applicability are validated through a real-world case study. Finally, actionable managerial insights are derived from comprehensive sensitivity analyses and tests.