Hybrid Optimization and Machine Learning Approaches for Enhanced 5G Network Slicing with Improved QoS and QoE

Network slicing is a critical technique in 5G networks that allows for the creation of multiple
virtual network instances that are customized to meet the needs of each customer. Each
slice functions as a distinct entity with its own unique attributes, including bandwidth and
Quality of Service (QoS). This feature enables the dynamic allocation of network resources
to optimize performance in accordance with the specific requirements of each user. In this
paper, we suggest a hybrid optimization and machine learning approach to improve the
slicing of 5G networks, thereby enhancing both the Quality of Service (QoS) and the Quality
of Experience (QoE). The methodology begins with a U-Net architecture optimized using
Whale Optimization (Whale-U-Net) for deep feature extraction from 5G network slicing
datasets. Subsequently, a hybrid meta-heuristic algorithm, Backtracking Search with
Quantum Newton Optimization (BS-QNO), is introduced to select optimal features. In order
to assess the influence of network slicing, we incorporate deep learning models, including
Long Short-Term Memory (LSTM) and Recurrent Neural Networks (RNNs), as well as
machine learning classifiers, including Decision Trees (DTs) and Support Vector Machines
(SVMs). Ultimately, the most desirable classification outcomes are optimized through the
implementation of a novel Deep Recurrent Reinforcement Learning (DRRL) methodology for
network slicing classification. Open-source benchmark datasets are implemented to verify
the proposed methodology. The results illustrate the efficacy of the proposed approach in
enhancing performance metrics and optimizing 5G network slicing.