Smart Routing: Applying Deep Learning for Efficient Network Traffic Management

The rapid expansion of modern networks in our current era, enhanced by the proliferation of connected devices with massive applications, has imposed challenges on network traffic management methods. In an expanding dynamic environment, the efficient use of network resources must be ensured while maintaining the reduction of latency, packet loss, and congestion. Deep Learning (DL) has emerged as an effective part of Artificial Intelligence (AI) and a technology capable of addressing the challenges by modeling complex patterns and adapting to network expansion. The proposed method is based on the neural network structure and exploiting feedback and back propagation in controlling the redesign of the neural network by calculating its weights and determining the priorities of extracted parameters. The variables are stored in vectors to be classified and find the best design to suit the network's adaptation to the dynamic expansion. The proposed model adapts to the network conditions and predicts appropriate paths in real-time. The good results prove the worth of the proposed model in terms of accuracy of 97.9% and prediction of 98%. The proposed model performs excellently in real time due to the dynamic ability to controlling the extracted parameters weight. In the future, hybrid algorithms based on machine learning and deep learning can be applied to obtain better results.