Procedure for designing a quasiorder on a set of multiattributed alternatives using reliable information about a decision-maker's preferences (Q1819706)

Using a set of performance-criteria based plant vector estimates, the design of binary preference-indifference relations that are, in the general case, a quasiorder, is studied. Basic requirements that must be met by solution methods assume the use of only the kind of information that the decision maker can easily provide based on data from descriptive investigations. The method must verify consistency of such information and must include an efficient procedure for eliminating possible contradictions. An algorithmic procedure is proposed for designing a quasiorder that meets these requirements based on an axiom that assumes independence of attributes by preference (to be checked) and transitivity of the desired relation.