Deep Reinforcement Learning for Multi-User Association in Fog Radio Access Networks


Fog radio access networks (F-RANs) are a promising evolution for future mobile communications. The evolution is a hybrid centralised-distributed architecture to only centralised cloud radio access networks (C-RAN), aiming to alleviate strain on the fronthaul links used for communication. The new architecture comes with challenges that cannot be tackled easily with the same algorithms used for solutions to problems raised by C-RAN. New efficient algorithms have to be developed to meet the requirements of the upcoming generation of mobile communications. Joint resource allocation and joint transmission are some of the ways that improvement is proposed, however they result in complex decision making processes, that are infeasible to solve for in real-time using brute force solutions. Reinforcement learning, particularly a recently introduced dual deep Q-network (DDQN) algorithm is acknowledged as a possible solution, however it has lengthy training times. This talk will present a novel distributed dual deep Q-network (3DQN) algorithm by introducing experience exchange in partially observable Markov decision process (POMDP) environments.


Jiangzhou Wang has been a Professor since 2005 at the University of Kent, U.K. His research interest is in the area of mobile communications. He is a Fellow of the Royal Academy of Engineering, U.K., Fellow of the IEEE, and Fellow of the IET. He was the Technical Program Chair of the 2019 IEEE International Conference on Communications (ICC2019), Shanghai, the Executive Chair of the IEEE ICC2015, London, and the Technical Program Chair of the IEEE WCNC2013.