Incorporation of the geographical positions of mobile users into the resource assignment process in uncoordinated heterogeneous cell deployments, can lead to significant improvements of cell and user throughputs. This paper proposes a novel algorithm that combines the knowledge of the users’ positions with a Q-learning and game-theoretic approaches to enhance the dynamic physical resource allocation during carrier aggregation (CA) in a semi-and uncoordinated deployment of Heterogeneous Networks (HetNet). The algorithm is evaluated through MATLAB simulation setup and in terms of macro-and pico- cell and user throughputs. It has been shown that regardless of the approach chosen for physical resource assignment, positioning information increases the system and user performances. Use of Q-learning and positioning information leads to increased cell throughput without degrading the user experience. Application of a game theoretic approach for resource assignment leads to increased fairness for both macro-and pico- users.

• Plamen Semov
• Vladimir Poulkov
• Albena Mihovska
• Ramjee Prasad

IEEE Wireless Communications and Networking Conference (WCNC) – Workshop on Interference and Design Issues for Future Heterogeneous Networks, 2014.

https://ieeexplore.ieee.org/document/6934858