Online Offloading for Energy-Efficient and Delay-Aware MEC Systems With Cellular-Connected UAVs

In this article, an unmanned aerial vehicle (UAV)-mobile edge computing (MEC) network is considered, where cellular-connected UAVs can either handle the computing tasks locally or offload to base stations. Considering the emerging computation-intensive and delay-sensitive applications, how to strike a balance between energy and delay, is one of the key design issues in UAV-MEC networks. Against this backdrop, we establish a double-queue model innovatively, in which the virtual queue is introduced to sense the backlog status of the actual queue. As such, the delay guarantee of computing tasks can be turned into the stable control of virtual queues. The network quality and server heterogeneity are considered to schedule the workloads rationally. Based on the Lyapunov optimization method, we formulate the deterministic problem to achieve a tradeoff between the long-term energy consumption and the time-average traffic delay, by jointly optimizing the offloading decision, resource allocation and trajectory planning, subject to the constraints of queue stability, resource budgets and flying kinematics. To solve this mixed-integer nonlinear programming problem, we propose an energy-efficient and delay-aware online algorithm, in which the problem is first split into equivalent resource allocation and trajectory planning subproblems, and the closed-form solutions of power, slot and computing resource allocation can be derived. Then, based on the Lagrange dual and successive convex approximation methods, these subproblems are solved iteratively to explore the optimality. Extensive simulations validate the superiority of our proposed algorithm over various benchmark schemes, showing its effectiveness in minimizing the energy consumption while simultaneously maintaining the low latency.