User Preference Oriented Service Caching and Task Offloading for UAV-Assisted MEC Networks
Ruiting Zhou, Yifeng Huang, Yufeng Wang, Lei Jiao, Haisheng Tan, Renli Zhang, Libing Wu
Abstract
Unmanned aerial vehicles (UAVs) have emerged as a new and flexible paradigm to offer low-latency and diverse mobile edge computing (MEC) services for user equipment (UE). To minimize the service delay, caching is introduced in UAV-assisted MEC networks to bring service contents closer to UEs. However, UAV-assisted MEC is challenged by the heavy communication overhead introduced by service caching and UAV’s limited energy capacity. In this article, we propose an online algorithm, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">OOA</i>, that jointly optimizes caching and offloading decisions for UAV-assisted MEC networks, to minimize the overall service delay. Specifically, to improve the caching effectiveness and reduce the caching overhead, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">OOA</i> employs a greedy algorithm to dynamically make caching decisions based on UEs’ preferences on services and UAVs’ historical trajectories, with the goal of maximizing the probability of successful offloading. To realize the rational utilization of energy from a long-term perspective, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">OOA</i> decomposes the online problem into a series of single-slot problems by scaling the UAV’s energy constraint into the objective, and iteratively optimizes UAV trajectory and task offloading at each time slot. Theoretical analysis proves that <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">OOA</i> converges to a suboptimal solution with polynomial time complexity. Extensive simulations based on real world data further show that <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">OOA</i> can reduce the service delay by up to 33% while satisfying the UAV’s energy constraint, compared to three state-of-the-art algorithms.