Sixth generation (6G) mobile networks promise unprecedented connectivity with ultra-high data rates, near-zero latency, and high reliability for applications like autonomous systems and extended reality. However, managing diverse resources, communication, computing, and caching (3C), poses significant challenges. Artificial intelligence (AI) is the key to automating and optimizing resource management in 6G. This review compares traditional model-based methods with adaptive data-driven approaches, covering computing and caching resource management alongside radio resources within the edge-cloud continuum. Crucially, the paper examines big AI models (BAIM) and the shift toward agentic AI for holistic, autonomous network automation. To address the opacity of these complex models, we highlight explainable AI (XAI) and digital twins (DTs) as an essential, combined trust and validation layer. Together, they ensure transparency, mitigate the computational overhead of real-time explanations, and enable safe training for critical functions like network slicing and multi-access edge computing (MEC) computation offloading. Finally, key challenges and future research directions for AI integration in next-generation wireless networks are outlined.
Building similarity graph...
Analyzing shared references across papers
Loading...
Rasini P. Amarasooriya
Mark A. Gregory
Shuo Li
Ad Hoc Networks
RMIT University
Building similarity graph...
Analyzing shared references across papers
Loading...
Amarasooriya et al. (Wed,) studied this question.
www.synapsesocial.com/papers/69e9b6aa85696592c86eb020 — DOI: https://doi.org/10.1016/j.adhoc.2026.104264