The Method Research For Multiple Attribute Group Decision-making Based On Multi-granularity Linguistic Explanation

In the multiple decision-making problems, the evaluation of attribute may use the qualitative form as language, just because of the limits of quantitative evaluation. In the aspects of group decision-making, different decision-maker use their own language evaluation sets, which called the problem of multiple attribute group decision-making based on multi-granularity linguistic explanation. This paper will based on the problem which are proposed, firstly, three different consensus methods to uniform the multi-granularity evaluation sets, and then, implement and improve the methods which used to integrate the language information is also given, mainly as follows:(1) A multi-granularity linguistic decision making model based on IOWA algorithm is proposed. In this model, the fuzzy sets transform method is used to uniform the multi-granularity linguistic, and then, based on the IOWA algorithm, the preference information provided by every decision maker is aggregated into group preference and most desirable alternative is selected. Finally, applied to the ERP selection, at meanwhile, an example is given to illustrate this method is effective and practicable.(2) Proposing the model for objective attribute weight making based on the multi-granularity language representation. In this method, firstly, a transformation function is given to uniform the multi-granularity evaluation matrix into the form of two-tuple linguistic information assessed in basic linguistic term set. Then, the method of TOPSIS is applied to ensure unknown attribute weight, based on the two-tuple aggregation operator and T-OWA operator selecting the most desirable alternative.(3) Research the method of attribute weight determination. Multi-granularity linguistic evaluation sets is introduced; by using the method of maximizing standard and mean deviations to ensure unknown attribute weight, evaluation scores are calculated through the hybrid aggregation (HA) operator. At the last, an actual example about risk investment is given to illustrate the practically and effectiveness of the proposed method.