Joint Partial Offloading and Resource Allocation for Parked Vehicle-Assisted Multi-Access Edge Computing

Abstract

In recent years, parked vehicle-assisted multi-access edge computing (PVMEC) has emerged to expand the computational power of MEC networks by utilizing the opportunistic resources of parked vehicles (PVs) for computation offloading. In this article, we study a joint optimization problem of partial offloading and resource allocation in a PVMEC paradigm that enables each mobile device (MD) to offload its task partially to either the MEC server or nearby PVs. The problem is first formulated as a mixed-integer nonlinear programming problem with the aim of maximizing the total offloading utility of all MDs in terms of the benefit of reducing latency through offloading and the overall cost of using computing and networking resources. We then propose a partial offloading scheme, which employs a differentiation method to derive the optimal offloading ratio and resource allocation while optimizing the task assignment using a metaheuristic solution based on the whale optimization algorithm. Finally, evaluation results justify the superior system utility of our proposal compared with existing baselines.