Agentic AI for Low-Latency Edge Caching and Resource Allocation in Internet of Vehicles

Abstract

Internet of Vehicles (IoV) is a key technology that enables comfort driving experiences and content-rich multimedia services, and the realization of IoV will lead to an unprecedented growth of the vehicular content. To deal with the massive content and provide real-time content delivery services, edge caching is proposed to deploy cache resource in road-side units or base stations to provide nearby caching and content delivery services. However, the complex multi-access vehicular communication, heterogeneous cache capability, and vehicular mobility pose great challenges for designing an efficient caching and content delivery scheme. To address these challenges, this paper proposes a novel Agentic AI-driven vehicular edge caching and resource allocation framework, which enables edge systems to autonomously perceive multimodal environments, perform contextual reasoning, and proactively adapt their decisions through continuous perception–reasoning–action loops. Specifically, the vehicular edge caching and resource allocation framework is proposed in which vehicles obtain the required content from local cache server or their adjacent cache servers. Then, the problem of vehicular edge caching and resource allocation is formulated to minimize the total content delivery latency. Then, we use Agentic AI to jointly optimize edge caching and resource allocation. Numerical results based on Uber dataset show that, compared with two benchmarks, the proposed algorithm can significantly improve the network performance and reduce delivery latency.