Robust optimization of emergency warehouse location and resource allocation in natural disaster-prone regions

Abstract

Earthquake-prone regions in western China face substantial emergency-logistics challenges due to sudden disasters, disrupted transportation networks, and highly uncertain material demand. To improve pre-disaster defensive planning and post-disaster response efficiency, this study develops a robust optimization model for emergency warehouse location, capacity selection, and material allocation. The model incorporates dual uncertainties–demand fluctuation and warehouse disruption–and integrates construction, transportation, and response-time costs within a unified budgeted-uncertainty framework that ensures tractability while guarding against worst-case conditions. A case study based on the 2025 Shigatse earthquake in Tibet shows that the demand-uncertainty budget exhibits a clear saturation threshold around $\Gamma _d \approx 8$, beyond which additional conservatism yields only limited marginal benefit. The results further show that robust optimization induces a defensively reconfigured deployment pattern relative to the deterministic benchmark, while policy-driven weight adjustment changes the intensity of conservatism without altering the existence of the saturation pattern. These findings provide quantitative support for designing protection-oriented and operationally feasible emergency material reserve systems in disaster-prone regions.