Optimal Scale Selections of Inconsistent Generalized Multi-scale Ordered Decision Information Systems with Multi-scale Decision Attributes Based on the Set-Valued Dominance Matrix
摘要
In many real problems, people often deal with ordered data problems with different hierarchies, and the optimal scale selection is also a hot issue of multi-scale information systems. This paper aims at the problem of optimal scale selection of an inconsistent generalized multi-scale ordered decision information system with multi-scale decision attributes (GMODIS for short). Each condition attribute and decision attribute of each object in GMODISs can take different attribute values at different scales. First, the concept of set-valued dominance matrix of GMODISs is proposed, and the condition for judging whether an ordered decision information system corresponding to a scale selection is consistent is given. Then, in an inconsistent GMODIS, the concept of generalized decision based on dominance relation is introduced. At the same time, a new consistent GMODIS is constructed by defining the dominance relation of interval value, and an algorithm for finding an optimal scale selection based on the set-valued dominance matrix of GMODISs is designed. Finally, the effectiveness of the proposed generalized decision optimal scale selections and the advantages of the set-valued dominance matrix for finding optimal scale selections method are verified by some numerical experiments.