Temporal and Spatial Spectrum Assignment in Next Generation OFDMA Networks through Reinforcement Learning

This paper proposes a dynamic spectrum assignment strategy in the context of next generation multicell orthogonal frequency division multiple access networks. The proposed strategy is able to dynamically find spectrum assignments per cell depending on the spatial and temporal distribution of the users over the scenario. Reinforcement learning methodology has been employed to implement the strategy, which compared with other fixed and dynamic spectrum assignment strategies shows the best tradeoff between spectral efficiency and quality-of-service.