A Markov decision process model for enhancing resilience in food supply chains during natural disasters

Abstract

Natural disasters like hurricanes, earthquakes, and floods devastate food supply chains and can threaten food security and public health. These disruptions, from production to consumption, lead to shortages, increased waste, and heightened vulnerability among food-insecure populations. This study addresses the need for effective emergency strategies to ensure food continuity and equity during crises. A Markov Decision Process (MDP)-based model is proposed to enhance food supply chain resilience under disaster conditions. The model involves a two-stage decision-making process: Stage 1 focuses on strategic decisions for immediate response, such as facility reconstruction, and Stage 2 handles tactical decisions during relief efforts, such as transportation routes and product flow. The objective functions of our model include minimizing response time and costs and ensuring equity of food accessibility. A resilience assessment approach is proposed to evaluate the performance of Pareto solutions. The proposed method is applied to the Qatar beef supply chain during a flooding scenario, demonstrating practical effectiveness. Sensitivity analysis is conducted to identify critical thresholds for establishing alternative distribution centers, which helps to optimize responses based on facility capacity. This research improves disaster preparedness and response, ensuring that food supply chains can adapt and recover quickly while enhancing the equity of people’s access to food and nutrition. A case study on Qatar’s beef supply chain under flood conditions shows that the proposed method achieves up to 95 % reduction in response time cost, a 9 % improvement in system resilience, and maintains over 99.5 % food accessibility under severe disruption scenarios.