Learning-Based Computation-Aware Access-Point Clustering and Joint Radio–Compute Allocation in Dynamic Cell-Free MEC Networks
Year: 2026
Type: conference-paper
Abstract: User-centric cell-free massive MIMO is a key 6G technology to meet strict latency and energy-efficiency demands of compute-intensive applications. In practical O-RAN deployments, the central unit and distributed unit compute capacity fluctuates with network-slicing dynamics. Relying on fixed compute resources and access-point (AP) clustering based solely on radio conditions leads to radio–compute mismatches, which in turn create bottlenecks that degrade both latency and energy efficiency. To address this drawback, we propose a communication and computation-aware AP-clustering strategy with joint resource allocation, leveraging a multi-agent reinforcement learning framework. The learned policies reduce average device energy consumption by 45% versus a fixed-compute RL benchmark and by 70% versus an offloading-first FPC heuristic across 10–50 Mbit/s, while maintaining 98–100% deadline satisfaction.
Institutions: Iquadrat (Spain), University of Patras
Language: English
Topic: IoT and Edge/Fog Computing
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7168248359