Study On Methods For Multiple Attribute Decision Making Under Some Situations

This paper has a deep going and systematic study on the multiple attribute decision making (MADM) problems, including the multiple attribute group decision making (MAGDM) problems, in which the attribute values take one of the forms of real numbers, interval numbers and fuzzy linguistic assessments. Four different priority methods are presented for fuzzy complementary judgement (FCJ) matrices, and a common formula for checking the consistency of the FCJ matrices and three approaches to improving the consistency of the FCJ matrices are given. The concept of hybrid judgement (HJ) matrix is defined, and a linear-objective programming method is presented for priorities of the HJ matrices. The priority theory of HJ matrices is then established preliminarily. The ordered weighted averaging (OWA) operator of Yager is extended to accommodate uncertain or fuzzy conditions, and a series of operators for aggregating decision information are given, including the ordered weighted geometric averaging (OWGA) operator, uncertain OWA operator, combined weighted geometric averaging (CWGA) operator, generalized induced ordered weighted averaging (GIOWA) operator, hybrid aggregating operator and hybrid linguistic weighted averaging (HLWA) operator, etc. Some methods based on these operators for MADM are proposed and then applied to solve some actual problems in the fields of supply chain management and venture capital. Some concepts, such as complex attribute ideal value of alternative, satisfaction degree of alternative, are defined, and a single-objective optimization model based on satisfaction degree of alternative is established. A two-step-method for the MADM problems is presented. Two linear objective programming models are established under the situations where the decision maker (DM) has preference information on alternatives, which takes the form of reciprocal judgement matrix and complementary judgement matrix respectively, and a method based on the models is presented for ranking alternatives in MADM. Motivated by the interactive idea of multiple objective decision making with infinite alternatives, the concepts of alternative achievement scale and alternative comprehensive scale are defined, and an interactive method based on the alternative achievement scale and the alternative comprehensive scale for the MADM problems is proposed. Some practical and effective methods are presented for the MADM problems in which the attribute values are interval numbers and the DM has (or has no) preference information on alternatives. These methods are then applied to the partner selection of virtual enterprises under some situations.