Intelligent Task Offloading in MEC ‐Enabled IoV Systems: From Data‐Driven Learning to Knowledge‐Based Reasoning

ABSTRACT

The rapid growth of the Internet of Vehicles (IoV) has created an urgent need for low‐latency, energy‐efficient and intelligent task processing in highly dynamic vehicular environments. This paper presents a comprehensive review of intelligent offloading approaches for MEC‐enabled IoV networks from an expert systems perspective. First, we systematically classify existing methods into four categories: heuristic‐based, optimization‐based, learning‐based and hybrid intelligent approaches. A comparative analysis is then conducted based on key performance metrics, including latency, energy efficiency, scalability, adaptability and sustainability. Beyond conventional performance evaluation, this review highlights the growing importance of knowledge‐driven decision‐making, explainability and hybrid reasoning mechanisms in enhancing the transparency, reliability and trustworthiness of offloading decisions in safety‐critical vehicular applications. Furthermore, the paper identifies key research challenges, including mobility awareness, real‐time adaptability, privacy preservation, interoperability and carbon‐aware edge intelligence. A conceptual framework integrating deep reinforcement learning with expert systems is also discussed to illustrate future directions toward explainable and sustainable offloading. Finally, open research opportunities are outlined, emphasizing the development of intelligent, adaptive and environmentally sustainable offloading frameworks for next‐generation 6G‐enabled IoV systems.