QoE-Aware Computational Resource Allocation for Connected Vehicles in Smart Urban Environments

Smart urban environments aim to enhance living standards by delivering effective and responsive services to residents. The growing number of connected objects, however, places increasing demands on computational efficiency and service provisioning. Leveraging the advancements in Information and Communications Technology (ICT), Connected and Autonomous Vehicles (CAVs) can serve as valuable computational resources to support service delivery. These vehicles can play the role of Vehicles as Computational Resources (VaCRs) by sharing their computational resources within smart cities. However, ensuring Quality of Experience (QoE) for service requesters, based on their diverse preferences, poses significant challenges in selecting and allocating resources. This paper presents a QoE-aware computational resource allocation system for Connected Vehicles (CVs), aimed at enhancing service delivery and computational efficiency in dynamic urban settings. The system models user requests based on key QoE factors and employs Performance Evaluation (PE) and QoE models developed using Multi-Criteria Decision-Making (MCDM) and machine learning techniques. A hierarchical multiagent architecture supports system deployment and coordination, while a QoE-aware game-theoretic model guides fair and efficient resource allocation. Compared to prior work, the proposed system demonstrates significantly improved performance in simulations, achieving higher classification accuracy (up to 96.5%) and lower average costs for service delivery. These results confirm the system’s effectiveness in harnessing vehicular computational resources and optimizing QoE in smart city environments.