Optimization of propagation prediction hybrid model for the cellular network design

A hybrid optimization model useful to predict the electromagnetic field in a cellular network is proposed. The proposed hybrid prediction model allows to obtain accurate results in considerable shorter calculation time than that necessary for the traditional ray-tracing. These methods mainly consist in a statistic prediction on limited extension areas, based on deterministic techniques. The proposed hybrid algorithm works in two different steps: the former calculates the electric field in a subset of the original observational points; the latter evaluates the power received in the remaining points by a statistic analysis of the values previously simulated. In this work we compare different methods such as linear regression, minimum mean square error and higher order regression curves to be applied in the statistic analysis in order to obtain more accurate results in the path loss calculation.