Demand-Aware Multi-Area Multi-UAV Empowered Mobile Edge Computing: A Joint Energy and Delay Optimization

Abstract

Multi-UAV-assisted mobile edge computing (MEC) has emerged as a promising paradigm for smart city infrastructures, where multiple uncrewed aerial vehicles (UAVs) collaboratively provide airborne computing services by bringing computational resources closer to end users, thereby reducing both communication costs and service latency. Existing multi-UAV-assisted MEC approaches typically employ fixed allocation schemes that predetermine the number of UAVs for a service area, failing to adapt to heterogeneous regional demands and leaving substantial UAV fleet capacity underutilized. To address this issue, we consider a multi-area multi-UAV-assisted MEC scenario where the control center jointly optimizes UAV deployment, user association, and resource allocation for each service area based on actual regional demands. This optimization enables multiple UAVs to be allocated to high-demand areas while assigning fewer UAVs to low-demand areas, thereby maximizing the service capacity of the entire UAV fleet. We formulate the optimization problem as a constrained multi-objective optimization problem (CMOP) that simultaneously minimizes the total system energy consumption and average user task completion delay, subject to deployment, user association, resource capacity, and task completion delay constraints. To effectively solve this CMOP, we propose a constrained multi-objective evolutionary algorithm that systematically reconstructs infeasible solutions into feasible ones through a constraint-guided solution reconstruction mechanism, thereby accelerating convergence toward feasible regions. Experimental evaluations demonstrate that our algorithm outperforms five state-of-the-art baseline methods in solution diversity and convergence performance. The results validate the effectiveness of flexible UAV allocation strategies, establishing a foundation for demand-aware resource provisioning in multi-UAV-assisted MEC systems.