Abstract
While several decision support systems (DSS) for quantitative problems with multiple attributes have been developed, approaches to DSS for qualitative problems are insufficient. In this paper, we present an effective evaluation method for DSS to measure decision maker's preference for qualitative attributes. The method provides utility values on a cardinal scale from decision maker's weighted paired comparisons among all of the alternatives. Since decision maker's weighted paired comparisons are usually inconsistent in a sence of cardinal transitivity, hypothetical weighted paired comparisons nearest to his or hers are found so that they satisfy conditions for cardinal transitivity. If his or her weighted paired comparisons are far from cardinal transitivity, the method suggests modification of them to be more satisfiable of cardinal transitivity. We present an applicational system of our method, which is used to support evaluation of some journals of computers.
