This work proposes an optimized scheme to deploy sensor networks. The network growth model is based on cellular automata to improve measurement quality at those locations likely to report sensor activity. Then, a population-based optimization scheme is performed to identify those solutions maximizing measurement quality while minimizing cost at every time step.

A Markovian formalism which includes a non-homogeneous term proportional to measurement is used as a prediction model for the computation of activity likelihood at a given location during a time interval. By numerical simulation it is found that, under suitable parameter conditions, the optimized solution outperforms the random deployment.