Improved Bat Algorithms for Solving Flowtime Minimization Task Scheduling Problems in D2D-Enabled MEC

Device-to-device (D2D)-enabled mobile edge computing (MEC) is a promising communication network in which computation tasks on mobile devices (MDs) can not only be processed locally but also can be scheduled to the MEC server or nearby MDs via D2D links in order to improve computational and energy efficiency. Due to the explosive growth of information and consequent large-scale data, the conflict between reducing task response time and the limited resources of MEC server and MDs has become an essential bottleneck. In this paper, we consider a D2D-enabled MEC system and study a task scheduling problem for minimizing the flowtime of all computation tasks under the energy constraints of MDs. To solve this problem, we develop a family of metaheuristic algorithms based upon our proposed framework of improved bat algorithm (IBA), which incorporates multiple best bat individuals. Also, we introduce a new linear updating scheme to update the bat positions and facilitate the evolution process of the metaheuristics. More importantly, we provide a rigorous theoretical analysis to prove the convergence of IBA by using this new updating scheme. Furthermore, we employ an improved initial solution generation strategy for enhancing the effectiveness of IBA. Numerical experiments demonstrate that the proposed algorithms can achieve high-quality scheduling solutions, in terms of shorter flowtimes, more effectively and more efficiently over baseline algorithms.