Abstract: This paper mainly focuses on this situation in the Multiple Criteria Decision Making(MCDM) problems with interval number that decision makers always have a basic expectation on some criteria. Firstly,we filter off the useless alternatives and find out the useful criteria that can arrive to the basic satisfaction degree of the decision makers on all criteria.By using the useful alternatives construct a new data table and select out the best alternative from the new table.Secondly,we propose a new definition of tolerance relation for criteria value and use the tolerance degree of criteria value to obtain the criteria weights. Thirdly,we also propose a new definition for satisfaction degree of alternatives.The best alternative(s)in decision making found by the traditional ranking methods does not have an equivalence relation with the alternative(s)that have the biggest combinatorial satisfaction degree of decision makers.Finally,an interval number example shows that the method is scientific and effective as well as feasible.

Key words: multiple-criteria decision making, expectation, satisfaction degree, filter off, tolerance relation, obtain criteria weights