Multiple obnoxious facility location: the case of protected areas

Most of the existing obnoxious facility location models use points to represent locations of demand centers and facilities. These points are typically centroids of geographical regions, such as cities or protected wildlife areas. Representing areas as points is convenient for the development and analysis of location models, but it may hinder the models’ practical applications. In continuous models, facilities may be located far from the centroid but close to (or even inside) the boundaries of the protected area. Also, some centroids may be located outside of the areas which they represent. We propose a heuristic that leverages the existing point-based models to locate multiple obnoxious facilities around protected areas and, at the same time, minimize the negative impact on those areas. The effectiveness of our approach is illustrated with both generated and real-world instances.