Computation Resource Configuration for Vehicular Edge Computing: A Fluid-Model Based Approach

Yue Zhou, Changyan Yi, Kun Zhu

Abstract: In this paper, the computation resource configuration for vehicular edge computing is studied. Unlike most of existing work in the literature, we focus on the problem of determining the optimal edge computing resource allocation to vehicles with computation offloading requests for maximizing the long-term management profit of the network operator (i.e., the road-side unit of the vehicular network) under the randomness of vehicular traffics and task processing. To describe the heterogeneities among different vehicles in terms of their edge computing service requirements and associated service payments, a multi-class management framework is employed. By considering the steady state nature of the system with dynamic evolutions, a novel fluid model is developed, which facilitates the formulation of the corresponding resource optimization problem. Based on this, we propose a simple but efficient solution, called threshold-based computation resource configuration scheme (TCRCS). Theoretical analyses and simulations evaluate the performance of the proposed solution and demonstrate its superiority compared to counterparts.