Physical property-aware multi-UAV spatial coordination for energy-efficient mobile edge computing

Abstract

In recent years, Unmanned Aerial Vehicle (UAV)-assisted Mobile Edge Computing (MEC) systems have emerged as innovative solutions for delivering efficient communication and computing services to Internet of Things (IoT) devices. However, the three-dimensional deployment and trajectory decision of UAVs remain challenging due to their highly non-convex and complex process characteristics. Existing methods often face scalability limitations, hindering their applicability to collaborative tasks as the number of UAVs increases. Furthermore, many approaches rely on simplified UAV models, neglecting the complexities of real-world physical dynamics. To address these issues, we propose a joint optimization framework designed to simultaneously minimize real-world UAV system overhead and enhance Air-to-Ground (A2G) communication capabilities. Our approach incorporates a deployment and trajectory design strategy that captures the comprehensive kinematic and dynamic properties of UAVs. In light of the problem’s inherent nonconvex structure and computational intractability, we introduce a collaborative multi-operator Differential Evolution (DE) variant algorithm with a semi-adaptive strategy, termed CSADE. This algorithm utilizes three distinct mutation strategies and integrates an external archiving mechanism to optimize both the number and locations of UAV Task Points (TPs). Additionally, we present an end-to-end dynamic UAV allocation and integrated flight path optimization method to ensure efficient route planning. The proposed method is evaluated through experiments on four data instances and compared with two related algorithms. Results demonstrate that our approach significantly reduces system operating costs while maintaining effectiveness and stability, highlighting its potential for large-scale UAV-assisted MEC applications.