Performance Enhancement using Q-Learning and Dynamic Clustering in Internet of Things Applications

A comprehensive framework termed QLDC-IoT, integrating machine learning and dynamic clustering to optimize connectivity and resource allocation in dense High Wireless Connectivity Network (HWCN) environments. Leveraging Extreme Gradient Boosting (XGBoost) classification models enhances handover decision-making processes in heterogeneous mmWave and LTE networks. Moreover, an adaptive clustering model using the K-Means Clustering effectively distributes resources and oversees connectivity in compact wireless networks. Through rigorous experimentation in the ns3 simulation environment, the QLDC-IoT framework demonstrates notable enhancements in data success rate, loss mitigation, delay reduction, throughput optimization, and routing efficiency, offering valuable insights for protocol design and implementation in diverse (Internet of Things) IoT applications. The research contributes to advancing IoT technologies, addressing challenges in connectivity, resource management, and network efficiency to support the future development of connected communities.