Event Details
Interference Mitigation in Multi-Tier Heterogeneous Network using Reinforcement Learning
Presenter: Abdullah Alqhtani
Supervisor: Dr. Fayez Gebali
Date: Wed, April 22, 2015
Time: 13:00:00 - 00:00:00
Place: EOW 430
ABSTRACT
ABSTRACT:
Efficient resource allocation in femtocell networks has become necessary owing to the enormous advantages of having femtocells deployed in a heterogeneous network. However, the interference arising from this deployment necessitate a mechanism for mitigation and optimal control of resource allocation. Q-learning, as an example of a reinforcement learning technique has been widely used for this purpose with more emphasis on downlink interference problems. Using a simplified heterogeneous model comprising of one macrocell and two femtocells, we extend the use of Q-learning to specifically model and address the uplink interference problem. We show by means of controlled experiments, that the proximity of the users in the network to their respective base stations, and the available power transmission levels plays a significant role in the total capacity of the network. This has the potential to enable networks to act in an improved manner. Results from the simulation can be used to configure any realistic network model.