TOC: Joint Task Offloading and Computation Reuse in Vehicular Edge Computing

With the proliferation of intelligent vehicles, addressing the demands of computing-intensive and delay-sensitive vehicle tasks has become a formidable challenge. Vehicle edge computing (VEC) has been proposed as an advanced paradigm that leverages edge servers such as road side units (RSUs) to offload tasks, thereby enhancing vehicle services. However, similar computation tasks in the VEC environment result in computational redundancy, imposing additional burden on the limited edge resources. Moreover, the increased interdependency among different tasks of vehicle tasks adds complexity to the offloading strategy. To this end, we propose a collaborative task offloading and computation reuse framework, called TOC, which enables RSUs to reuse previous computations and design a task offloading scheme based on a Conflict Graph(CG) model. We also evaluate the efficiency and effectiveness of TOC using real-world datasets, and our results show that TOC is able to reduce the task completion time by 48.73% compared to baselines.