| Popis: |
Ojien, when several information sources are available, data are both heterogeneous and asynchronous. The combination of all these information sources remains as a difjicult task, which strongly depends on the representation of the used data. Consequently, it is imperative to choose a model of knowledge representation well adapted to each kind of information. When each source is pe$ectly represented and modelled, we need to know how to associate them the most faith/ul and the most reliable way. In this paper, we propose an algorithm multi- criteria association using belief theory for resolve the problems mentioned above. This association algorithm will provide a more representative synthesis of the whole knowledgeprovided by a set of fiili or partial expert advices built from heterogeneous data. Once the initial belief masses were built pom the multi-criteria association. it is possible to combine them with a multi-object association algorithm. |
