A hidden Markov random field approach for control prioritization based on noisy observations
2016
Bonneau, Mathieu | Martin, Julien | Romagosa, Christina | Johnson A., Fred | Unité de Recherches Zootechniques (URZ) ; Institut National de la Recherche Agronomique (INRA) | United States Geological Survey (USGS) | University of Florida [Gainesville] (UF)
Control of invasive species is receiving increasing attention, both in terms of the development of methodological tools and the documentation of negative effects of such species on native flora. A recurrent question focuses on the problem of optimal spatial allocation of resources in order to minimize the invasive population, while respecting practical constraints. Several theoretical frameworks exist to compute the optimal management strategy when the spatial distribution of the species is known; however, knowledge on spatial distribution is not often available. We propose to use a method from the field of image analysis, where the problem of images segmentation from noisy data is common and several statistical methods are available. We adapted these methods to the management of invasive species where only a rough estimate of the invasive spatial distribution is known. This first estimate is then used to compute the probability of presence of the species over space. These probabilities are in turn used to set-up an integer linear programming problem to compute the optimal allocation of control effort under cost constraints. We illustrate our method with the example of melaleuca (Melaleuca quinquenervia) control.
Afficher plus [+] Moins [-]Mots clés AGROVOC
Informations bibliographiques
Cette notice bibliographique a été fournie par Institut national de la recherche agronomique
Découvrez la collection de ce fournisseur de données dans AGRIS