With the increase of vehicular applications, which aimed at improving traffic safety and driving experience, the limited computing resources of vehicles face the challenge of the real-time task requirements. Vehicular Edge Computing is an efficient way to augment the computational capabilities of vehicles. However, there are few researches have simultaneously considered the vehicle reliability, task prioritization, and timely scheduling of tasks. To bridge this gap and motivate vehicles to share computing resources, this paper proposed a multi-task offloading method. We formulated a revenue maximization problem and designed a control algorithm for this problem based on deep reinforcement learning. Meanwhile, we employed blockchain technology to ensure vehicle reliability. Simulation results validate the effectiveness of the proposed method, demonstrating a profits improvement of 67.98%.
Multi-Task Computation Offloading based on Vehicular Blockchain
17.11.2023
415838 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch