A bi-criteria approach for the data association problem
Hugot, Hadrien; Vanderpooten, Daniel; Vanpeperstraete, Jean-Michel (2006), A bi-criteria approach for the data association problem, Annals of Operations Research, 147, 1, p. 217-234. http://dx.doi.org/10.1007/s10479-006-0069-9
TypeArticle accepté pour publication ou publié
Nom de la revueAnnals of Operations Research
MétadonnéesAfficher la notice complète
Résumé (EN)The data association problem consists of associating pieces of information emanating from different sources in order to obtain a better description of the situation under study. This problem arises, in particular, when, considering several sensors, we aim at associating the measures corresponding to a same target. This problem, widely studied in the literature, is often stated as a multidimensional assignment problem where a state criterion is optimized. While this approach seems satisfactory in simple situations where the risk of confusing targets is relatively low, it is much more difficult to get a correct description in denser situations. This is why, we propose, for the first time to our knowledge, to address this problem in a multiple criteria framework using a second complementary criterion, based on the identification of the targets. Due to the specificities of the problem, simple and efficient approaches can be used to generate non-dominated solutions. Moreover, we show that the accuracy of the proposed solutions is greatly increased when considering a second criterion. A bi-criteria interactive procedure is also introduced to assist an operator in solving conflicting situations.
Mots-clésMultidimensional assignment problem; Data association; Bi-criteria optimization; Interactive procedure
Affichage des éléments liés par titre et auteur.