ADAPTIVE MACHINE LEARNING-DRIVEN RESOURCE ORCHESTRATION FOR ENERGY-SUSTAINABLE WIRELESS NETWORKS
Keywords:
energy efficiency; wireless resource allocation; reinforcement learning; traffic prediction; edge computing; O-RAN; concept drift; QoS; 5G; 6G.Abstract
The rapid growth of wireless infrastructure and heterogeneous traffic increases the need to jointly optimize network performance
and energy consumption. Static allocation, analytical optimization, and independently trained learning policies can become difficult
to maintain when channel conditions, traffic demand, queues, and computational workloads change over time. This paper presents
AML-RA, an adaptive machine-learning-driven resource orchestration framework
References
[1] “Deep Reinforcement Learning Method for Energy Efficient Resource Allocation in Next Generation Wireless Networks,” ACM
ICCNIoT, 2020, DOI: 10.1145/3398329.3398332.
[2] “Energy-Efficient Subchannel and Power Allocation for HetNets Based on Convolutional Neural Network,” arXiv:1903.00165.
