Robust optimization of pharmaceutical emergency logistics network under dynamic demand

This study aims to optimize pharmaceutical emergency logistics under dynamic demand and disrupted routes during public health crises. By integrating multi-scenario analysis and multimodal transportation, it seeks to minimize response time, unmet demand penalties, and costs while balancing efficiency and equity. The model addresses limitations of traditional single-mode logistics, leveraging COVID-19 case data to enhance adaptability in resource allocation. A robust optimization model is developed, integrating dynamic demand forecasting, scenario probabilities, and capacity constraints across four epidemic stages. The NSGA-III algorithm is employed to solve multi-objective trade-offs, with performance compared against NSGA-II using metrics like spacing and Pareto ratio. Robust standard vectors and scenario probabilities are analyzed to evaluate stability, supported by computational experiments from Chinese cities like Wuhan. NSGA-III outperformed NSGA-II, generating 60% more Pareto solutions in T4 with 3% faster computation. Robust vectors significantly influenced outcomes: γ3 increased penalty costs linearly in high-demand phases, while γ1 escalated procurement expenses over time. Scenario probabilities p3 reduced penalties by 15–20% through coordinated logistics. The framework enables emergency managers to prioritize air transport for urgent deliveries and establish centralized hubs, reducing average response times by 18%. Public-private partnerships and dynamic inventory adjustments improve equity and efficiency, particularly in high-risk regions. This study contributes to the field by unifying dynamic demand modelling, multimodal transport optimization, and robust scenario-based decision-making into a single analytical framework. The application of NSGA-III effectively resolves many-objective optimization challenges, outperforming traditional methods in both diversity and convergence. A scenario-driven parameter analysis is introduced to quantitatively assess the impacts of uncertainty, thereby advancing theory in crisis logistics management.