WSEAS Transactions on Information Science and Applications
Print ISSN: 1790-0832, E-ISSN: 2224-3402
Volume 22, 2025
CRLSA: Cognitive Reinforcement Learning Spectrum Allocation in Addressing Spectrum Scarcity in 5G Wireless Communication
Authors: , , , , ,
Abstract: With the development of 5 g wireless communication networks, spectrum allocation (especially in metropolises) is becoming increasingly challenged. To mitigate this problem, in this paper, we propose a Cognitive Reinforcement Learning Spectrum Allocation (CRLSA) framework to improve spectrum utilization while satisfying the quality of service (QoS) requirements. Deep Q Networks (DQN), an advanced deep-reinforcement learning technique, serve as the backbone of the framework, complemented with challenging urban spectrum management features. So it uses DQN to obtain effective spectrum allocation policy in simulated urban areas. Using synthetic datasets, spectrum management with DQN agent-based dynamic resource allocation takes into account available spectrum bands, QoS metrics, interference levels, and user mobility patterns to optimize performance. We build an environment to collect data from the previous step by simulating the environment during the training phase in which the agent learns the knowledge and skills he needs to make good decisions about spectrum allocation. In addition, CRLSA framework configuration and optimization are critical to improve its performance. The framework is tuned with hyperparameter adjustments, reward shaping, and exploration strategies to enable better convergence and effectiveness in real-world deployment scenarios. Moreover, computations are optimized to guarantee real-time decision-making in changing urban surroundings. CRLSA necessitates analysis and a better comprehension of the communications
channels beyond interference, including propagation conditions, and acknowledges that the overall performance is contingent on whether multiple, possibly conflicting objectives are realistically harmonized.
Search Articles
Keywords: 5G Wireless Communication, Spectrum Allocation, Cognitive Reinforcement Learning, Deep Q- Networks (DQN), Urban Environments, Quality of Service (QoS)
Pages: 450-465
DOI: 10.37394/23209.2025.22.37