The Studying Of The Priority Methods Of Decision-making Matrix And Its Practicing

The question of MADM generally exists in the sphere of decision-making and it has extensive practical background. The arrangement methods of decision-making matrix are the important components of MADM. Rich achievements on this aspect of theoretical study have been made and many decision-making methods have been presented. But the studying and practicing about the arrangement methods of decision-making have just begun, therefore the theoretical study on this aspect is yet imperfect and necessary to be strengthened. Only for this reason, in this thesis, the author probes into the arrangement methods of decision-making from the both aspects of attribute weights known and attribute weights completely unknown.In the first chapter, the author introduces the basic concept on MCDM and the related contents of MADM, sums up the studying survey of the arrangement methods of decision-making matrix among MADM, and summarizes the main content and significance of this thesis.In the second chapter, the author puts forward several arrangement methods of decision-making matrix when attribute weights are known. When the attribute value is accurate number, two new arrangement methods are put forward. One is the improved TOPSIS method, which extends and reforms the normalized method of decision-making matrix used in TOPSIS, further simplifies the calculation of ideal alternative and negative-ideal alternative, and proposes one new easy-calculating relative closeness which is equivalent to the one defined by TOPSIS; the other is the projection method in ideal alternative. In both methods, the composite value of each alternative is not needed to calculate and its arrangement results can be obtained. Their advantages are concise, reasonable and clearly-thought. It is also easy to be dealt with by computers. When the attribute value is interval number, three arrangement methods are put forward in this chapter. One is the arrangement method of decision-making matrix based on deviation degree of interval number, the other two are the arrangement methods of decision-making matrix based on possibilitydegree of interval number. Finally some related calculating examples are given to demonstrate their effectiveness and feasibility of the several above-mentioned methods.In the third chapter, the author puts forward few methods of decision-making matrix arrangements when attribute weights are completely unknown. There also exist two respects when attribute values are accurate number and interval number. Several arrangement methods of alternative are proposed from the optimized perspective when weights information is completely unknown. These methods are mainly through establishing optimization model. By way of solving optimization model, attribute weights and alternative arrangements then can be obtained. When attribute value is accurate number, three methods are put forward, the first is the model constructed with the distance of ideal alternative; the second is the model constructed with relative closeness of ideal alternative; the third is the model constructed by utilizing the deviation between the vector of alternative integrated attribute value and the vector of alternative ideal integrated attribute value. When attribute value is interval number, a model based on deviation degree of interval number and a method in which attribute weights can be obtained by utilizing the method of entropy are put forward. In the end a related calculating example is also given.