Research On A Class Of Complex Multiple Attribute Decision Making

As an important component of the modern decision-making science, the multi-attribute decision-making is widely used in many fields such as engineering design, economic systems, management and military etc. With the rapid social and economic development, and the rapid progress of science and technology in various fields, decision making problems become more and more diverse and complex. How to make more effective multi-attribute decision making has been an important issue so that people pay attention to conduct a deep research of it.In this paper, a class of unknown weights multiple attribute decision making for several common methods will be explored and studied, including the followings:First of all, it introduces the hybrid multi-attribute decision making problems, and explains the relevant knowledge of multi-attribute decision making with the attribute values being single point, interval and linguistic assessment information. It also introduces several multi-attribute decision making methods. An ideal-point based method is proposed where the attribute weights are unknown and the attribute values are the uncertain linguistic variables. In the proposed method, the decision matrix is normalized by defining the ideal-point and calculating the relative distance between linguistic attribute values and the ideal-points. A mathematical programming model is set up to determine the attribute weights by maximizing the overall values of the alternatives, and accordingly the overall values of the alternatives are obtained and their ranking is figured out. The simplicity and practicality of the proposed method are illustrated by comparing the method which is based on two-tuple linguistic.Secondly, it also studies some concepts about precision number, interval number, triangular fuzzy number of hybrid multiple attribute decision making problem. For multiple attribute decision making problems with interval uncertain variables and the unknown attribute weight, it proposes a method based on gray relational grade. The idea of the method is to construct a new relative closeness degree by combining the ideal-point and the gray correlation. The decision matrix with uncertain information is normalized by calculating the relative distances between the attribute values and the ideal points. Afterwards, a mathematic model is set up aiming at maximizing the overall values of the alternatives, and then work out the attribute weight values. Ultimately, the alternatives are ranked according to their overall values. A numerical example is provided to illustrate the proposed method. Its practicality and feasibility are illustrated by comparing the method which is based on entropy technology with the method which is based on gray relational grade.