Efficient and reliable hybrid deep learning-enabled model for congestion control in 5G/6G networks

Future generation networks such as millimeter-wave LAN, broadband wireless access systems, and 5th or 6th generation (5G/6G) networks demand more security, low latency with more reliable standards and communication capacity. Efficient congestion control is considered one of the key elements of 5G/6G technology that allows the operators to run various network instances using a single infrastructure for a better quality of services. Artificial intelligence (AI) and machine learning (ML) are playing an essential role in reconfiguring and optimizing the performance of a 5G/6G wireless network due to a vast amount of data. A smart decision-making mechanism is required for the incoming network traffic to ensure load balancing, restrict network slice failure and provide alternate slices in case of slice failure or overloading. To circumvent these issues, a hybrid deep learning-enabled efficient congestion control mechanism is proposed in this paper. This hybrid deep learning model consists of long short term memory (LSTM) and support vector machine (SVM). The applicability of the proposed model is validated by simulating for one week using multiple unknown devices, slice failure conditions, and overloading conditions. An overall accuracy rate of 93.23% is calculated for the proposed hybrid model that reflects the applicability. Apart from this, other performance metrics such as specificity, recall, time consumption, varying training, test sets, true-false rates, and f-score were used for the performance evaluation purposes of the proposed model.