Predicting cellular network performances with cell-level monte-carlo simulations
B. Pasquereau, Yann Le Helloco · 2006
In order to implement changes and optimize settings in the cellular networks they operate, radio planners are facing the challenge of validating those changes prior to their implementation, in order to minimize risks and costs. To do so, they have traditionally relied on the prediction capabilities of their design software tools, which implement propagation models and network analytical models. With the progression of technologies however, and especially the introduction of cognitive radio, analytical models are no longer capable of providing accurate results, so statistical modeling techniques have to be introduced. We present in this paper a new approach to system- level Monte-Carlo simulations, that draws interference patterns rather than user locations, as in traditional approaches currently used for CDMA systems, which leads to fast computation times while maintaining accuracy. We also present how this is applied to simulating performances of GSM/EGPRS networks, and present results obtained on real networks.