Dynamic Resource Allocation with Reinforcement Learning for 5G RAN Slicing

Information from the abstract

The continuous increase in mobile subscribers, coupled with the growth of diverse applications, directly impacts the demand for wireless communication. As 5G deployments occur globally, industry and academia are developing concepts and technologies to support 6G and beyond networks to meet this continuously surging demand. One of the core technologies for future communication networks is the development of techniques for efficiently managing dynamic frequency usage. Ultra Dense Networks (UDN) have been the primary network type deployed. The three main application classes have been specified for 5G networks, differentiated by different service performance requirements, and the network slicing technique has been proposed to divide 5G system resources, recognizing the different needs of various services. This work presents a technique to improve the efficiency of managing a limited frequency spectrum dynamically using the Reinforcement Learning (RL) technique. The approach takes into account the needs and usage patterns of wireless communication services. Results from the numerical analysis show that the system performance when implementing the proposed technique has improved with suitable configuration parameters compared with systems operated with traditional resource allocation approaches in terms of system throughput, spectrum efficiency, packet loss rate, and fairness, and the performance is satisfactory for individual traffic measurements.

Why this record is monitored

This record has an Impact Signal of 71/100 based on recency, source, collaboration, and bibliographic signals. It prioritizes monitoring and is not a judgment of research quality.

Related topics: Software-Defined Networks and 5G · Advanced Wireless Network Optimization · Cognitive Radio Networks and Spectrum Sensing