Intelligent Optimization of Hybrid Networks: Decision-Making with Integrated Approaches

The rapid expansion of heterogeneous networks necessitates sophisticated optimization techniques to enhance the overall network performance. Traditional network optimization approaches, such as the per-layer and narrow scope of cross-layer optimization, often fail to capture global network dynamics because of their limited scope. In this study, we propose a comprehensive integrated optimization framework that utilizes a neural network with a decision-making algorithm to optimize gateway selection in hybrid networks comprising local 5G (L5G) and Wi-Fi. Our approach is validated through simulations in EXata, demonstrating improvements in normalized through-put compared to random- and RSSI-based selection strategies. Quantitative results show that our proposed ML-based method enhances network performance by up to 13% compared to the RSSI-based method in dense network scenarios, suggesting its potential utility in next-generation networks.