Load change detection and energy saving in wireless networks

Contact: Ramin Khalili

We have data collected from 96 cells by an operator from a real world cellular network in Switzerland. The data captured SMS and call loads of these cells over a period of one week.

Our goal is to study and propose algorithms that detect changes in distribution of cell loads and react to these changes by switching on or off some of the base stations in the network (to save energy).

Our method would be based on nonparametric on-line quality control procedure approach proposed by David McDonal et al. (some introductory slides can be found here). A good knowledge about performance evaluation (in particular, different prediction and forecasting methods) is necessary.