Research On Problems For Group Decision Making Based On Interval Numbers Gathering

Multi-attribute group decision making is an important extension of the direction of the management decision problems. With the development of society, the problems people encountered are becoming more and more complex, and those properties are becoming more and more. Just depending on personal power alone cannot easily solve such a large and complex issue so. As a result, multi-attribute group decision making becomes more and more popular, and becomes the key issue of discussion. Multi-attribute group decision-making problems are basically discussed by a way of facing group evaluation. The evaluators score to evaluate problems through some representative attributes. These attributes constitute a decision matrix, and then use some specific methods to process the information matrix in order to find out the optimal solution.Now with the people in-depth study of the multi-attribute group decision-making problems, it is discovered that the experts often have a big difference in view of things according to their respective capabilities, preferences and experiences when scoring on the property of the questions. It cannot fully meet the need of the experts when using the real number as scoring approach. They believe that things are often not the absolutely determinate, many of which are fluctuated within a certain range, so that using interval numbers to score the property index should be more appropriate and accurate. Multi-attribute group decision-making problem in this paper is to study the form of interval numbers to describe the property index for the background. In the whole of the property completely unknown, information gathers by a specific method to evaluate the pros and cons of the various options value and finally the optimal solution will be found.It is an important part to gather the rational and effective information on group decision-making problems. Gathering information has a decisive impact on the ultimate decision-making problems. This article points out the study of the group decision-making problems for interval information gathering aspect is insufficient at this stage, and proposes the idea for a new way of interval number information gathering. Using the principle that interval numbers become a stable presence in the form of coordinates of points in two-dimensional space to establish the non-linear programming model and using plant growth simulation algorithm for generalized “Fermat point” for solving optimization problems. Finally, compare the difference between the optimal aggregation method and the simple weighting method(IVWA), and display its accuracy.After the interval number information gathered, it will need some ways to sort their advantages and disadvantages. Through the establishment of the method combining optimal the clutch point and projection pursuit model to solve the multi-attribute group decision-making problems. First introduce projection pursuit model based on accelerating genetic algorithm, and then elaborated on how to couple the best clutch point nonlinear programming model and projection pursuit model. The specific method is that let the interval number get in the normalization process, then information gathering, and obtain the optimum clutch-point decision-making preferences matrix. The matrix is then placed in the PPC model based on accelerating genetic algorithm to determine the optimal projection values, which are sorted to obtain the optimal solution.In order to verify the feasibility and accuracy of the model, the paper gives a example of a large-scale data which has a multi-attribute group, and ultimately get the result of the sort and the best solution. The author then applies the method to actual cases of virtual enterprise partner selection and gets the final results. Through the contrast with other relevant literature and the comparative analysis, the results are consistent with the results of comparative literature, and ultimately reach the same conclusion which illustrates this method not only has high feasibility and effectiveness, but also more simple and flexible compared to the comparative literature. It can be applied to actual investigation of virtual enterprise partner selection problem.