Find in Library

The advent of the Internet of Vehicles (IoV) has sparked strong scholarly interest in the determination of dependable offloading destinations for tasks. The lack of global information, however, prevents the existing research from being applied in the context of information asymmetry. This paper proposes an original framework for intermediary vehicle-assisted task offloading (IVATO) in scenarios of information asymmetry. Through IVATO, we introduce an intermediary vehicle election mechanism fully based on trust and information mastery. Specifically, a new method is developed to evaluate trust based on resilience. In addition, we conceive a degree of information mastery to measure the amount of information in the vehicle. Based on the intermediary vehicle selected, an objective function for designing the offloading strategy is formulated to maximize both reliability and effectiveness. Proximal Policy Optimization (PPO) based deep reinforcement learning algorithm is adopted to tackle the optimization problem. Simulation results show that the proposed trust evaluation method is more rational than the existing methodologies in the long term. The proposed offloading mechanism shows a 49% increase in utility and a 45% increase in reliability over other schemes.