Study On Group Decision Making Approaches Based On Uncertain Fuzzy Judgment Matrixes And Ordinal Preference Information

The complexity of groups and group decision-making is an expression form of the complex system in complexity science. The realization of group behavior is getting through the individual competition and mutual cooperation. The group behavior has the characteristics of instability, nonlinear, uncertainty, fuzziness, unpredictability, and so on. Multi-criteria decision-making and group decision-making with uncertain preference have been appeal fields in recent years. However, the theories and methods of group decision-making with uncertain information is far from being perfect and mature. There have few models and methods of group decision-making problems with fuzzy or incomplete preference information.In this thesis, multi-criteria group decision-making problems with uncertainty fuzzy judgment matrixes (such as interval number judgment matrix, linguistic judgment matrix, triangular fuzzy number judgment matrixes, trapezoidal fuzzy number judgment matrix, etc.) and group decision-making problems with ordinal preference information have been researched. At last, a case study of the group decision-making evaluation on sustainable development potential of larger metal resource bases is performed.The main research content and the innovations are as follows:(1) A new AIP (Aggregation of Individual Priorities) group decision-making method with decision-makers' weight based on interval judgment matrixes is proposed. This method bases on the interval criteria value. It mixes the "sum-product algorithm" with the interval number operation rules, and the decision-makers' weight is being taken into account. And then the individual priorities are aggregated by "improving relative entropy model". This method avoids the loss of decision makers' information, and it has solved the problem of the inconsistency of group preference. (2) A new "two-tuple linguistic" group decision-making method based on linguistic judgment matrixes is proposed. For a group decision-making problem with linguistic criteria value based on the multi-granularity linguistic term sets, this method puts forward an extended EOWA operator—T2-EOWA operator. And it applies the T2-EOWA operator to turn the individual judgment matrix into consistent judgment matrix. And it achieves the last result by aggregating of the consistent judgment matrixes. This method solves the inconsistency of multi-granularity linguistic judgment matrixes in group decision-making, and it also avoids the loss of decision makers' linguistic information. It ensures the reasonability and the effectiveness of decision results.(3) This thesis analyses the relationship of the Classification Quality (parameter-γ) and the Inclusion Degree (parameter-β) based on the Variable Precision Rough Set Model, and it constructs an algorithm to ascertain the parameter-βarea based on given parameter-y value. Through ascertaining the parameter-β, the ability of adapting noisy data of decision tables can be improved.Based on the relationship of parameter-y and parameter-β, a new group decision-making method is proposed. This method determines the relative importance of attributes according to the training data in decision table, and then it constructs and aggregates the judgment matrixes. These judgment matrixes meet the demand of consistency. This method can handle the data with noise, and it can analyze and deal with the inaccurate, inconsistent and incomplete information.(4) A new group decision-making method based on different types of judgment matrixes with incomplete information is proposed. Three different conditions of judgment matrixes have been considered in this method. The first, the types of preference information in each judgment matrix is different. There has real number judgment matrix, interval judgment matrix, triangular fuzzy number judgment matrix, or trapezoidal fuzzy number judgment matrix. The second, the characteristic of each judgment matrix is different. There has complementary judgment matrix, or reciprocal judgment matrix. The third, the judgment matrixes are all with incomplete information. The consistency is also been considered. There has formed the more complete methodology to solve group decision-making problems as above.(5) Two new group decision-making methods based on ordinal preference information are proposed. Based on ordinal preference information, the group vote matrix can be constructed. Then the group vote matrix is converted to group complementary judgment matrix (methodⅠ) or group reciprocal judgment matrix (methodⅡ). And the judgment matrix will be converted to a derivative consistent matrix by mathematical methods. The final sorting results can be obtained through solving the derivative matrix. This method can avoid the "draw results" which always appears in the Social Selection Function Methods. This method can ensure the rationality of decision-making results.