A Q-learning-based dynamic channel assignment technique for mobile communication systems

This paper deals with the problem of channel assignment in mobile communica tion systems, In particular, we propose an alternative approach to solving the dynamic channel assignment (DCA) problem through a form of real-time re inforcement learning known as Q learning. Instead of relying on a known tea cher, the system is designed to learn an optimal assignment policy by direc tly interacting with the mobile communication environment. The performance of the Q-learning-based DCA was examined by extensive simulation studies on a 49-cell mobile communication system under various conditions including h omogeneous and inhomogeneous traffic distributions, time-varying traffic pa tterns, and channel failures, Comparative studies with the fixed channel as signment (FCA) scheme and one of the best dynamic channel assignment strate gies (MAXAVAIL) have revealed that the proposed approach is able to perform better than the FCA in various situations and capable of achieving a simil ar performance to that achieved by the MAXAVAIL, but with a significantly r educed computational complexity.