Optimization of resilient humanitarian logistics using a robust combinatorial multi-attribute reverse auction

Abstract

Effective procurement of relief items is critical for mitigating the impact of disasters and ensuring timely assistance to affected areas. However, existing humanitarian procurement models often fail to address multi-attribute, multi-source, and multi-period structures under uncertainty, limiting their practical effectiveness. This study presents a two-period combinatorial multi-attribute reverse auction framework leveraging multiple procurement strategies including pre-positioned inventory, spot markets, backup suppliers, and reverse auction to enhance the resilience and mitigate disruption risks of humanitarian logistics. The proposed model spans two periods, each involving bid construction and bid evaluation for the reverse auction. For each supplier, the bid construction problem determines optimal order quantities, bundle price, and delivery time. The bi-objective bid evaluation problem is optimized using a fuzzy multi-objective method. Key features of this work include substitution options, partial fulfillment policies, and prioritization of relief items based on urgency. To cope with uncertain parameters, robust optimization is employed. The model is validated using a case study from District 4 of Tehran, using 2024 population data reported by the Tehran Municipality, demonstrating its effectiveness in minimizing costs and response times while ensuring supply chain continuity. The results emphasize that reinforcing local distribution centers enhances responsiveness during disasters, while engaging backup and spot market suppliers strengthens flexibility under disruption. Moreover, effective budget allocation significantly reduces procurement costs and delivery delays, ensuring supply continuity and improved resilience across the humanitarian logistics network.