Facility Location and Routing Optimization for Humanitarian Drone Logistics

Abstract

Indonesia is vulnerable to natural disasters, which frequently disrupt ground transportation infrastructure and hinder humanitarian aid distribution. In such conditions, drone-based logistics offers a promising alternative for last-mile delivery to affected areas. This study proposes an integrated optimization framework for humanitarian drone logistics that combines a multi-objective Facility Location Problem (FLP) and a Vehicle Routing Problem (VRP). The FLP determines the optimal locations of drone stations to maximize service coverage while minimizing infrastructure and operational costs, and is formulated as a min-max weighted-sum objective with normalized components. The VRP, solved using a Simulated Annealing (SA) algorithm, optimizes pickup-and-delivery routes with multi-trip capability, subject to drone capacity, distance, and operational time constraints. The framework is evaluated using real disaster data from the 2022 Cianjur earthquake in Indonesia. Results show that a small number of strategically located drone stations achieve high service coverage at a reasonable daily cost, with weight coverage approaching completeness. Routing optimization substantially reduces the deployed fleet relative to the FLP upper bound, demonstrating the added value of the VRP stage as a cost-reduction mechanism. The findings demonstrate that integrated strategic and operational optimization significantly enhances the efficiency and feasibility of drone-based humanitarian logistics, and that objective weight tuning can serve as a practical policy lever for balancing coverage, cost, and equity in disaster relief operations.