Computation Offloading for Latency Reduction in Regionalized Hierarchical Vehicular Fog Network

Regionalized communication and computation architecture aggregates idle vehicular fog computation resources in urban cities without dedicated infrastructure deployment. Vehicular fog computing migrates computation services close to the edge of the network, which can break the long response latency limitation of cloud computing. However, the high mobility and random distribution of vehicles lead to the difficulty of the fog infrastructure deployment and the offloading decision. In this paper, we investigate the latency minimization problem in the regionalized three-tier vehicular fog network, which is divided into the sub-problems of single-region offloading decision and multi-region resource allocation. In particular, we make the preliminary offloading decision for users in each roadside unit (RSU) coverage area to solve the single-region offloading decision problem. We also transfer the multi-region resource allocation problem into a group-knapsack problem to obtain the optimal resource allocation results among all the users in the base station coverage area. Simulation results validate the superiority of our proposed algorithm, which enables latency reduction in the regionalized three-tier vehicular fog network.