Reinforcement Learning and Q-Learning for Resource Allocation in MIMO Network with Intelligent Reflective Surfaces
Jasim Khudhair Salih Turfa (Electrical and Computer Engineering College, Altinbas University, Istanbul, Turkey), Oguz Bayat (Department of Electrical and Electronics Engineering, Yeditepe University, Istanbul, Turkey)
Abstract
This research invokes a hybrid technique based on combining two intelligent algorithms of R-learning and Q-learning processes applicable to modern wireless communication systems based on IRS and MIMO networks. R-learning determines whether actions are desired according to signal quality criteria, while Q-learning localizes the user through generating a criterion based on minimum BER response. The combination achieves a significant reduction in BER of the received signal, approaching the ideal case without noise or fading effects. An analytical study comparing increasing IRS elements versus algorithmic solutions shows significant enhancement in signal quality with increasing IRS elements.