Research Article
An Improved Dynamic Spectrum Access Algorithm Based on Reinforcement Learning
@INPROCEEDINGS{10.1007/978-3-031-30237-4_2, author={Chen Zhong and Chutong Ye and Chenyu Wu and Ao Zhan}, title={An Improved Dynamic Spectrum Access Algorithm Based on Reinforcement Learning}, proceedings={Machine Learning and Intelligent Communication. 7th EAI International Conference, MLICOM 2022, Virtual Event, October 23-24, 2022, Proceedings}, proceedings_a={MLICOM}, year={2023}, month={4}, keywords={Dynamic spectrum access Reinforcement learning Q-Learning Cognitive radio networks}, doi={10.1007/978-3-031-30237-4_2} }- Chen Zhong
Chutong Ye
Chenyu Wu
Ao Zhan
Year: 2023
An Improved Dynamic Spectrum Access Algorithm Based on Reinforcement Learning
MLICOM
Springer
DOI: 10.1007/978-3-031-30237-4_2
Abstract
This paper proposes an improved dynamic spectrum access algorithm based on reinforcement Learning in cognitive radio networks. Q-learning algorithm is used as the core to update the optimal strategy for the established Markov decision process according to specific scenarios, and Q-table is updated iteratively to improve the learning rate. In order to verify the effectiveness of the proposed algorithm, we construct the mathematical model and the simulation environment. The simulation results validate the effectiveness of the proposed algorithm, which can effectively improve the system throughput under the condition that not affect primary users’ communication. The proposed algorithm can quickly adjust the corresponding reward value and strategy in the iterative process of reinforcement learning training, so as to quickly converge to the optimal strategy, and the training results are consistent with the expected results.
