AINeutralarXiv – CS AI · 8h ago6/10
🧠
Optimization-as-a-Service via Multi-Agent Large Language Model for Radio Access Networks
Researchers propose a multi-agent large language model system to optimize physical resource block allocation in 6G radio access networks, treating optimization as a service that dynamically adapts to real-time network conditions. The framework uses a closed-loop architecture with scene understanding, objective generation, and reflection agents, achieving near-optimal performance with minimal inference latency through a novel one-shot distillation mechanism.