The Research Of Multiple Attribute Error-Eliminating Decision-Making Method

Multiple attribute decision making (MADM) problem which had attracted a lot of experts’attention, is widely exist in every aspect of people’s life and production. The decision making psychology of avoiding error loss which is widely exist in the human decision processes, was proved by a lot of experiments. Therefore, it is reasonable and necessary to research the multiple decision making problem in the view of decreasing and (or) avoiding error-loss. Error-eliminating theory is a new theory in which error is took as research object, the decreasing or avoiding error-loss are took as the research purpose. How to use error-eliminating theory to solve multiple attributes problem is obviously worth to research. In this paper, the decision making method which use error-eliminating theory to solve multiple problem is called multiple attribute error-eliminating decision-making (MAEEDM) method, and crisp-important type, fuzzy (interval)-important type, linguistic-important type, hybrid-key type and double hybrid type these five kinds of multiple attribute decision making methods are detailed Study.Firstly, the theoretical base about the research of multiple attribute error eliminating decision making method is given. The concepts of error, error function, error system and so on are given on the basis of error-eliminating theory. These concepts are used to extend out the concepts of error, key attribute, important attribute, redundant attribute, attribute error value, attribute error-limit-loss value and so on in the situation of multiple attribute decision making. Error function and error-loss function about key attributes and important attributes, the main steps and thoughts.Secondly, three kind of important type multiple attribute decision making method are studied respectively.(1) For the crisp important multiple attribute decision making problem, the corresponding error is given and the way of computing attribute error-limit-loss value is given too according to the projection principle. Three decision-making methods as error-eliminating summation, error-eliminating centrifugation, and error-eliminating TOPSIS are given. These three methods are combined by the method of Copeland and a compound decision-making model is built.(2) According to the steps and process of crisp important multiple attribute decision making method, attribute error function for interval number is given and interval number is used to describe comprehensive error-loss value of alternative strategies. The possibility degree method is used to rank the comprehensive error-loss value. And the fuzzy (interval)-important multiple attribute decision making method is built at last.(3) For the character of linguistic information, the related error function is given. The linguistic evaluation information is changed into error value which is expressed in the form of2-tupe linguistic. The ways of computing attribute error-limit-loss value and attribute error-loss value are given according2-tuple linguistic analysis method. Error-loss series of alternative strategies are derived by using the2-tuple linguisitic algorithm. The aggregated approach of error-loss series is given, and and the alternative strategies were prioritized and selected accord to the aggregated result.And than, the hybrid key multiple attribute error-eliminating decision making method is studied. For the hybrid key multiple decision making problem, the concepts of non-optimal and non-inferiority strategies are given, the classic error function of some kind of key attributes is built. The reason why the attribute error value is equal to attribute error loss value is given. The non-optimal strategy is screened out by the form of computing comprehensive error value. The non-inferiority strategies are prioritized and selected by the number of error deviation degree.At last, double hybrid multiple attribute decision making problem is researched on the basis of3kinds of important multiple attribute decision making method and hybrid key multiple attribute decision making method. The attribute reduction method is given, for the reason of the redundant attribute without any influence in the process of decision making. The ways of computing the comprehensive error-loss values of key attributes and important attributes by referring to another four kinds of multiple attribute decision making method, which have been research in the above. At last, strategies are prioritized and selected by error-loss2-tuple which is built in the paper.