IoT-5G and B5G/6G resource allocation and network slicing orchestration using learning algorithms

Abstract

The advent of 5G networks has precipitated an unparalleled surge in demand for mobile communication services, propelled by the advent of sophisticated wireless technologies. An increasing number of countries are moving from fourth generation (4G) to fifth generation (5G) networks, creating a new expectation for services that are dynamic, transparent, and differentiated. It is anticipated that these services will be adapted to a multitude of use cases and will become a standard practice. The diversity of these use cases and the increasingly complex network infrastructures present significant challenges, particularly in the management of resources and the orchestration of services. Network Slicing is emerging as a promising approach to address these challenges, as it facilitates efficient Resource Allocation (RA) and supports self-service capabilities. However, effective network segmentation implementation requires the development of robust algorithms to guarantee optimal RA. In this regard, artificial intelligence and machine learning (ML) have demonstrated their utility in the analysis of large datasets and the facilitation of intelligent decision-making processes. However, certain ML methodologies are limited in their ability to adapt to the evolving environments characteristic of 5G networks and beyond (B5G/6G). This paper examines the specific challenges associated with the evolution of 5G and B5G/6G networks, with a particular focus on ML solutions for RA and dynamic network slicing orchestration requirements. Moreover, the article presents potential avenues for further research in this domain with the objective of enhancing the efficiency of next-generation mobile networks through the adoption of innovative technological solutions.