Several Typical Types Of Methods About Multiple Attribute Decision Making

The multiple attribute decision making problems exist everywhere. Such as investment decision, project evaluation, scheme optimization as well as filling volunteer text and so on. But because of the complexity, uncertainty of the problem and fuzziness of human understanding, which results in the value measurement or evaluation result of the multiple decision making problems can be exact number, interval number,language words, fuzzy numbers, random variables and so on. For the three types which include the exact number, interval number and random variables, the attribute puts forward different methods of decision-making.In view of the attribute value of multiple attribute decision making is exact number, this article gets through the original method of multiple attribute decision making based on maximum entropy principle and maximum deviation method, using the solution of the multi-objective optimization problem to improve it.And puts forward a new multiple attribute decision making model based on the principle of the maximum entropy.According to the attribute values in multiple attribute decision making is interval number, this paper gives the concept of the relative satisfaction and the concept of the similarity, uses the principle of relative entropy to give an improved combination of weight determination method and proposes a interval multi-attribute decision-making method based on the similarity.For the attribute values in multiple attribute decision making is normal random variable,this paper gives two decision-making methods.The first decision making method is using the knowledge of probability and statistics to process to turn normal random variable to the interval number,use the interval multi-attribute decision-making model for decision making based on improved TOPSIS method.Another decision method is according to the relationship between the expectation and variance of different normal random variable to give and determine the decision theorem and proof of the random dominant relationship between them,and then by using the theorem on different attributes to determine the random dominant relationship between different schemes and establish the corresponding stochastic dominant matrix.Finally make the decision using ELECTRE III method.