Intelligent OpenRAN orchestration assisted by a Reinforcement Learning Resource Management policy
Wagner Silveira, Fábio Henrique Cabrini, Filippo Valiante Filho, Alberico C. Barros Filho, Sérgio Takeo Kofuji · 2021
Efficiently orchestrating resources in a Radio Access Network poses a challenging task, which has been made even more difficult by the addition of vertical slices of operation with very unrelated requirements in 5G and Beyond 5G. The proposal of an open and interoperable standard by the Open Radio Access alliance has paved the path to foment research and development of innovative solutions. As a result, the solution herein presented employs a Machine Learning algorithm of Reinforcement Learning to deal with the multiple slices requirements in a unified approach based on a resource management policy. The simulation results showed a consistent response of the Machine Learning model to wisely administrate the available resources to maximize the potential of the User Equipment requirements attainment.