Search

Due to the rapid development of the Internet of Vehicles (IoV), the combination of IoV and edge computing, known as vehicle edge computing (VEC), has received considerable attention from both academia and industry. However, task offloading in diverse intersection scenarios remains suffering from inefficiency of resource allocation and low quality of service for task execution due to the imbalance of traffic flow and the rigid requirement of latency. To address these issues, we develop a task offloading strategy by a fuzzy decision-making algorithm to handle uncertainty and imprecision. This task offloading strategy comprises two components: (1) A VEC resource pool with available vehicles at each intersection is constructed when taking the rotating direction of the recognition region. Then, we introduce a fuzzy decision-making algorithm to select a set of high-quality service vehicles from this VEC resource pool as an auxiliary edge server (AS). (2) We employ an edge service provider (ESP) to manage the computational resources of a main edge server (MS) and an AS deployed at a traffic intersection. The negotiation between the ESP and the task vehicles is modeled as a Stackelberg game. We prove the existence of the unique perfect Nash equilibrium, and a genetic algorithm is applied to find the optimum. Finally, we conduct simulation experiments with datasets collected in real-world scenarios. The results demonstrate that our scheme decreases task execution time by 9.73% compared to the cloud server scheme and reduces energy consumption by 13.78% compared to the state-of-the-art reinforcement learning (RL) strategy. [ABSTRACT FROM AUTHOR]

The emergence of vehicular edge computing (VEC) has introduced a new computational paradigm for high-quality processing of computing services in Internet of Vehicles (IoV) scenarios. However, due to the limited computational resources of the VEC server, it is not sufficient to adequately meet the demand for highly concurrent computational services in high-density vehicular communication networks. To address this issue, we consider an idle mobile vehicles assisted vehicular edge computing framework and propose a hybrid Stackelberg-Match cooperative task offloading and resource pricing algorithm (SMOP). The algorithm considers the mobility of vehicles and the duration of channels, coordinating the computational resources of fixed VEC servers and idle mobile vehicles within the vehicular network (VN). This enhances offloading efficiency and maximizes participant benefits. Specifically, the Stackelberg game is used to derive differentiated pricing schemes for idle mobile vehicles and VEC servers for different vehicular tasks, and the stable matching method is employed to determine task offloading strategies. Finally, we conduct experiments on a real Chengdu traffic dataset. The results demonstrate that the proposed solution effectively reduces offloading costs and exhibits strong robustness in handling latency-sensitive and data-intensive service requests. [ABSTRACT FROM AUTHOR]