Decision-making Models And Methods Based On Granular Computing

Fuzzy set, rough set and formal concept analysis are proposed by Zadeh, Pawlak and Wille R., respectively. All of them are effective tools for processing uncertainty problems in real life. Using the theories of fuzzy set, rough set and formal concept analysis, this paper investigates multigranulation decision-theoretic rough set in ordered information system, multigranulation soft rough set models, double-quantitative decision-theoretic rough set models and granular computing approach to two-way learning based on formal concept analysis in fuzzy datasets. The main innovations are as follows:1. From the perspective of granular computing, the dominance classes are handled with the partition function, and transform the non-probability measure into a probability measure space. Partition function is applied to the multigranulation decision-theoretic rough set in an ordered information system, and three kinds of this multigranulation decision-theoretic rough set model in an ordered information system are constructed. And a case study about how to make decisions according to the proposed models is introduced.2. From the perspective of granular computing, two types of multigranulation soft rough sets are firstly presented, which are optimistic multigranulation soft rough set and pessimistic multigranulation soft rough set. Then we investigate the relationship among classical soft rough set, optimistic multigranulation soft rough set and pessimistic multigranulation soft rough set.3. From the perspective of granular computing, two kinds of Dq-DTRS model are constructed, which essentially indicate the relative and absolute quantification. After further studies to discuss decision rules and the inner relationship between these two models, an illustrative case study about the medical diagnosis is introduced to interpret and express the theories.4. From the perspective of granular computing, a novel model of two-way learning system in fuzzy datasets is developed. Relationship between an object and its fuzzy attributes is discussed in view of information granules in this two-way learning system. We suggest how to learn necessary fuzzy information granules, sufficient fuzzy information granules, and necessary and sufficient fuzzy information granules from any fuzzy information granule.