Dynamic enhanced Inter-Cell Interference Coordination using reinforcement learning approach in Heterogeneous Network

This paper investigates enhanced Inter-Cell Interference Coordination (eICIC) techniques for Heterogeneous Networks (HetNets), and models this strategic coexistence as a multi-player system in which interference management strategies inspired from a form of reinforcement learning known as distributed Q-learning are devised. Specifically, this paper focuses on time domain eICIC techniques in which each macrocell optimizes its Almost Blank Subframe (ABS) configuration consisted of ABS density and reduces power level. Instead of relying on predefined configuration, the system is designed to learn optimal ABS configurations by directly interacting with the environment. To substantiate our theoretical findings, system level simulations are carried out in which our proposed solution is compared with the conventional approach that ABS configuration is fixed. Our proposed solution is shown to yield substantial gains of user throughput compared to fixed ABS configuration.