The Stochastic Dynamic Postdisaster Inventory Allocation Problem with Trucks and UAVs

• DOI: 10.1287/trsc.2023.0438
• Publication Date: 2025-02-27T15:43:58Z
• Authors: R. M. van Steenbe...; W. J. A. van Hees...; M. R. K. Mes
• Keywords: FOS: Computer and information sciences; Artificial Intelligence (cs.AI); Computer Science - Artificial Intelligence; Optimization and Control (math.OC); I.2.8; 1. No poverty; FOS: Mathematics; 93E35 (Primary) 90B50, 68T05, 68T20 (Secondary); Mathematics - Optimization and Control

Humanitarian logistics operations face increasing difficulties due to rising demands for aid in disaster areas. This paper investigates the dynamic allocation of scarce relief supplies across multiple affected districts over time. It introduces a novel stochastic postdisaster inventory problem (SDPDIAP) with trucks and unmanned aerial vehicles (UAVs) delivering goods under uncertain supply demand. The relevance this humanitarian lies importance considering intertemporal social impact deliveries. We achieve by costs (transportation deprivation costs) when allocating supplies. Furthermore, we consider inherent uncertainties areas potential use cargo UAVs enhance operational efficiency. study proposes two anticipatory solution methods based on approximate programming, specifically decomposed linear value function approximation (DL-VFA) neural network (NN-VFA) effectively manage process. compare DL-VFA NN-VFA various state-of-the-art (e.g., exact reoptimization proximal policy optimization) results show 6%–8% improvement compared best benchmarks. provides performance captures nonlinearities problem, whereas shows excellent scalability against minor loss. From practical standpoint, experiments reveal that consideration improved both Finally, deploying can play crucial role goods, especially first stages after disaster. reduces transportation together 16%–20% maximum times 19%–40% while maintaining similar levels demand coverage, showcasing efficient effective operations. History: has been accepted Transportation Science Special Issue TSL Conference 2023.