Designing a post-disaster relief supply chain: variable fixing-based heuristic

A well-coordinated post-disaster relief supply chain is essential for delivering timely assistance to affected regions under uncertainty. This study addresses network planning in the post-disaster phase by explicitly considering supply capacity and demand uncertainty within a robust optimisation framework. We propose a mixed-integer nonlinear programming model that integrates procurement, facility location, processing, and distribution decisions, while accounting for social costs such as deprivation and logistics. To efficiently solve this complex model, we develop a two-phase heuristic that combines nonlinear programming relaxation with a structure-preserving approach. Computational experiments evaluate the algorithm’s performance across different network sizes, service levels, demand variability, and social cost scenarios. A case study based on the 2008 Sichuan earthquake demonstrates real-world applicability. Benchmark comparisons with Genetic Algorithm and Particle Swarm Optimisation show that our approach provides superior solution quality and computational efficiency, especially for large-scale, uncertain supply chains. These results highlight the potential of robust optimisation and the proposed heuristic to improve disaster response planning.