A bi-objective optimization approach for wildfire detection

Abstract

We consider the problem of buying and locating equipment for covering a (discretized) region. We propose two approaches, based on mathematical programming modelling and the epsilon-constraint method, that allow obtaining the efficient frontier of a bi-objective optimization problem. In the first approach, the objectives are maximizing coverage and minimizing cost. In the second approach, lexicographic optimization is used to incorporate additional objectives - maximizing double coverage and minimizing the maximum fire rate of spread of uncovered points. The latter objective comes from the specific application that motivated this work: wildfire detection. We present results from a case study in a portuguese landscape, as an example of the potential of optimization models and methods to support decision making in such a relevant field.