The New Research On Ranking Methods Of Interval Numbers

The ranking methods of interval numbers are one of the basic problems in uncertain Multiple Criterias Decision Making. And it is also an important topic of uncertain Decision Analytic. The paper has made a study of the ranking methods of interval numbers. Mainly works are shown as follows:1. Point out that the ranking methods, which is based on interval correlative coefficient, have the same distances between some interval numbers and comparing datum, dued to the choosing of reference interval numbers and some other reason. It leads to the fact the method is difficulty to distinguish these interval numbers. Pointed to this shortcoming, a new concept so-called" the best and the worst reference interval as a datum intervals are proposal. And relative correlative coefficient as a ranking index measure interval numbers' priority degree is also brought in.2. Based on a new distance measure, a new similarity degree as ranking index of interval numbers is proposal. Interval numbers priority relationship educed by the new similarity degree accords with that of traditional similarity degree, but the new similarity degree has better distinguish ability than the old one.3. Point out that for some interval numbers, the priorities measured by some possibility degree definition are not delicate enough, and the consistency of the fuzzy complementary matrix built by these possibility degrees cannot be guaranteed. For covering the shortages, an indirect comparing idea is adopted, namely comparing with two interval numbers by means of comparing with the same reference interval respectively. Combining the idea, a new concept of priority degree to measure how one interval number is superior to another is brought in. The new definition can overcome possibility degree's shortcomings at a certain extent and reduce errors.