Joint caching and computing resource allocation for task offloading in vehicular networks. Issue 21 (5th February 2021)

• Record Type: Journal Article
• Authors: Wang, Zhi; Hou, Ronghui
• Is Part Of: IET communications. Volume 14:Issue 21(2020)
• Page Start: 3820
• Page End: 3827
• Publication Date: 2021-02-05
• Subjects: cellular radio -- cloud computing -- optimisation -- resource allocation -- cache storage -- linear programming -- mobile computing -- nonlinear programming -- integer programming; joint caching -- computing resource -- future vehicular networks -- computation intensive -- efficient content caching strategy -- computation resource allocation strategy -- mobile edge computing servers -- MEC -- caching content -- maximising caching performance -- caching capacity -- balanced caching strategy -- efficient caching strategy; Telecommunication systems -- Periodicals; Speech processing systems -- Periodicals; 621.38205
• DOI: 10.1049/iet-com.2020.0100
• Languages: English
• ISSNs: 1751-8628

Abstract

Abstract : To meet the requirement of constrained delay and computation resource of the future vehicular networks, it is imperative to develop efficient content caching strategy and computation resource allocation strategy in mobile edge computing (MEC) servers. In the proposed network framework, since the caching capacity and computing resource of each MEC are limited, and the coverage areas of MECs are overlapped, the vehicular networks have to decide what contents to cache, how to offload tasks and how much computing resource needs to be allocated for each task. In this study, in order to jointly tackle these issues, we formulate caching strategy, offloading decision and computing resource allocation coordinately as a mixed integer non‐linear programming (MINLP) problem. To solve the MINLP problem, we divide it into two subproblems. Firstly, we investigate a balanced and efficient caching strategy based on similarity in vehicular networks. Secondly, we apply McCormick Envelopes to convert MINLP problem into LP problem, and then adopt improved branch and bound algorithm to obtain the optimal offloading decision and computing resource allocation strategy. Simulation results indicate that the proposed schemes have a good performance in reducing economic cost under the deadline of each task.