An Intelligent Framework for Network Slicing to Enhancing Resource Scheduling in 5G Network

As the usage of Smart devices has increased exponentially due to the increased population, need for the high speed and efficient communication infrastructure is badly required. The 5 G network played a crucial role in supporting these applications as it affords well-organized solutions in meeting the varying requirement of each environment. Network Slicing plays a catalyst in building the intelligent and low latency 5G communication by managing the network and bandwidth efficiently, it also enables more efficient network resource scheduling. But, enhancing the performance of 5 G networks using slices remains to be major bottleneck for the researchers. To solve this research challenge, this research article proposes the novel Multi-variant Siamese Gated recurrent networks networks for achieving the best network performance. The framework was evaluated using the Kaggle 5G Slicing Datasets which was developed using tensorflow-keras python libraries. Nearly 400 K records which consist of 5 G dynamic features such as packet loss, packets transmitted and dates are used to construct the dataset environment and comprehensive experimentation is carried out to measure the performance metrics like accuracy, precision, recall, specificity, F1-score, latency and throughput. To prove the excellence of the proposed model, different deep learning models are analysed. Simulation results demonstrates that the proposed model has achieved the $\text{9 8. 6 \%}$ of slicing accuracy, $\text{9 7. 9 \%}$ of precision, $\text{9 7. 8 \%}$ of recall, 97.65 % of F 1 -score and excellence than the other models. Furthermore, the proposed model proves its efficacy by classifying the slices to the requested services with the reduced latency and bandwidth usage and enhance resource scheduling of 5G networks.