Technical Note: \cal Q-Learning
Machine Learning
Introduction to Reinforcement Learning
Introduction to Reinforcement Learning
Formalizing Multi-Agent POMDP's in the context of network routing
HICSS '03 Proceedings of the 36th Annual Hawaii International Conference on System Sciences (HICSS'03) - Track 9 - Volume 9
Cooperative Multi-Agent Learning: The State of the Art
Autonomous Agents and Multi-Agent Systems
BLISS '08 Proceedings of the 2008 Bio-inspired, Learning and Intelligent Systems for Security
A POMDP-based spectrum handoff protocol for partially observable cognitive radio networks
WCNC'09 Proceedings of the 2009 IEEE conference on Wireless Communications & Networking Conference
Game-theoretic deployment design of small-cell OFDM networks
Proceedings of the Fourth International ICST Conference on Performance Evaluation Methodologies and Tools
Femtocells: Technologies and Deployment
Femtocells: Technologies and Deployment
POMDP-Based Coding Rate Adaptation for Type-I Hybrid ARQ Systems over Fading Channels with Memory
IEEE Transactions on Wireless Communications
Decentralized cognitive MAC for opportunistic spectrum access in ad hoc networks: A POMDP framework
IEEE Journal on Selected Areas in Communications
Self-organized femtocells: a Fuzzy Q-Learning approach
Wireless Networks
| Hi-index | 0.00 |
We present a self-organized downlink power control for interference management when Home eNodeBs (HeNBs) work in co-channel operation with the macrocell system. The main novelty with regards to previous works is that we provide a completely autonomous framework, considering 3GPP release 11 hypothesis of non availability of X2 interface between evolved NodeBs (eNBs) and HeNBs. In this situation, the HeNB has to make autonomous decisions without receiving any feedback from the macro network. We model the HeNBs as a multiagent system where each node is an independent agent able to learn through Reinforcement Learning (RL) techniques a downlink power allocation policy for different interference situations. To deal with the lack of information in the scenario, we rely on the theory of Partially Observable Markov Decision Process (POMDP). POMDP works on the basis of a set of beliefs that the HeNB builds considering the impact it causes to the macrocell system. To gather this system performance information, we propose that HeNBs use spatial interpolation techniques, such as ordinary Kriging. Results show that the proposed approach allows HeNBs to autonomously learn a power allocation policy to coexist with the macro network, in a 3GPP compliant fashion, and without introducing overhead signaling in the system.