Collaborative Computation Offloading in the Multi-UAV Fleeted Mobile Edge Computing Network via Connected Dominating Set

In these years, Unmanned Aerial Vehicle (UAV) is widely employed as a flying server into the Mobile Edge Computing (MEC) network to carry stochastic computation task offloaded from mobile users. However, currently, much efforts are made on the individual computation capability of each UAV during the offloading task, while neglecting the coordination advantage of multiple UAVs in a fleeted way, leading to low efficiency on computing massive bursty data. In this work, therefore, we propose a twofold computation offloading mechanism in a collaborative way for a multi-UAV assisted MEC network using the Connected Dominating Set (CDS). In particular, we solve a system energy efficiency-maximizing optimization problem by using a well-tailored Alternating Direction Method of Multipliers (ADMM) algorithm and Lyapunov optimization. In the proposed mechanism, we develop a two-stage computation task partitioning strategy and a user scheduling scheme with optimized UAV trajectory via designing a CDS virtual backbone network. The numerical results indicate that the superiority of our proposed mechanism compared with the benchmark algorithm.