CC On/Off Scheduling Using Learning-Based Prediction for LTE in the Unlicensed Spectrum

In this letter, we propose a novel component carrier on/off scheduling scheme using the learning-based prediction for LTE in the unlicensed spectrum. Particularly, we use Lyapunov optimization to improve the performance of packet throughput and energy consumption, and use Vapnik-Chervonenkis theory to analyze the gap between the predicted optimal value and the practical one. The simulation results show that the proposed scheme can deliver good trade-off between packet throughput and energy consumption by turning off some carrier components, especially when the traffic load is not high.