Integrating UAVs and AI for Resilient Smart Energy Systems in Disaster

Natural disasters pose serious challenges to smart grid infrastructure by simultaneously disrupting power and communication systems, leading to isolated microgrids or &ldquoislands&rdquo. To support post-disaster recovery, this paper presents a multi-objective optimization framework for the efficient placement of unmanned aerial vehicle base stations (UAV-BSs) as mobile relay nodes connecting power sources (PSs) and static base stations (SBSs). The proposed method jointly optimizes geographical UAV-BS positions considering some conflicting objectives: minimizing the number of UAV-BSs and unserved PSs (NAPS), maximizing system throughput and energy efficiency. An enhanced Multi-Objective Reinforcement Learning (MORL) with clustering-based initialization is developed to improve convergence and solution diversity. Simulation results on the Simbench dataset confirm the effectiveness of the proposed approach in achieving robust, energy-efficient, and cost-effective UAV-BS deployment for smart grid restoration.