Task Classification for Optimal Offloading and Resource Allocation in Vehicular Edge Computing

Vehicular Edge Computing (VEC) aims to extend cloud service provision to the network edge vehicles, thus reducing back-haul network traffic and service latency. The emergence of computation-intensive tasks in connected, cooperative, and autonomous vehicles is inevitable. Vehicular applications and services generate time-sensitive tasks and require near real-time processing and updates. Executing such computationally intensive tasks on vehicles either leads to high computation latency or unavailability of resources. Recently, VEC has evolved as a popular choice for offloading and used for offloading these complex tasks. However, recent works put an upper limit on task offloading for execution. Moreover, executing a small task over aVEC server causes increased energy consumption and communication latency in a highly dynamic topology. Furthermore, due to the limited computational resources and capacity of the VEC server, it cannot process all tasks when the number of requests is high. Therefore, it is required to have an optimal selection for tasks offloading to solve this problem. A task-centric offloading and optimal resource allocation scheme based on task priority has been proposed in this paper. The proposed approach increases the probability of execution of these tasks on the VEC server as compared to other computation-intensive tasks. The numerical results demonstrate that the proposed work not only minimizes the processing delay but also has increased the percentage of tasks offloaded to the VEC server having higher priority in comparison to baselines.