Towards an interactive query environment for a multi-attribute decision model with imprecise data and vague query conditions

A multi-attribute decision model aiming at the selection of an efficient set of alternatives is presented. The major challenge is that the alternatives are not perfectly known, i.e., the associated values may be imprecise. Furthermore, for improved user friendliness, vague query conditions are allowed. Imprecise scores are described by probability distributions, vague query conditions by fuzzy predicates. The philosophy of the presented approach is to create an environment similar to the multidimensional spreadsheet front-end of data warehouses, in which the user can interactively discover the imperfect answers. In particular, numeric measures that allow an aggregated view of alternatives are defined. A geographic information retrieval scenario demonstrates the use of the decision model.