Enterprise femtocell network optimization based on neural network modeling

In future B3G/4G communication systems, the special application of femtocell in the enterprise offices and public places has broad prospect. However, as the femtocell operation under such multi-femtocell environment is significantly different from that of usual residential femtocells, the femtocell configuration and optimization might be much more complex. One key issue is to find how will the femtocell network performance be influenced by the change of the femtocell access point´s (FAP´s) parameters (power, assigned channel, e.g.) in the multi-FAPs environment. This paper proposes a neural network (NN) modeling approach to approximate the relation between the femtocell network performance and the FAPs´ operating parameters. Simulation results show that the proposed model works well and is very close to the actual situation.