Research On Methods Of Multiple Attribute Decision-Making And Integration

Multiple attribute decision-making, which is one of the most important problems of decision-making science, is widely used in many fields, such as: investment decision making, item evaluation, optimization alternatives, plant site selection, resources assignment, ranking of industrial department, bidding, and comprehensive evaluation on economic benefit, etc. It is greatly significant to study multiple attribute decision-making theory and approaches as well as their applications.In actual decision making, the integrated methods are appeared as the decision-making methods are restricted in application, and the problems are complicated. This paper investigates mainly integrating decision-making methods, and presents some new methods for the integrated problems.Due to the complexity of object things and fuzzy characteristics of human thought, in decision making problems, many indexes are not only denoted by real numbers, but also denoted by interval numbers and fuzzy numbers. In real life, the hybrid sequences including two or all of the three types of numbers is familiar, which is an integrated problem, and this paper investigates the methods of the multiple attribute decision making with integrated problem, and puts forward a new method for attribute weights.This paper introduces the methods to formalize the indexes, the attribute weighting methods and the classical methods of decision-making.As to the defects of subjective and objective weighting methods, a new weighting method is proposed, and the new model is established, which is based on the least square theory by using weight adding and relational analysis, and the more rational weight values are computed.Aiming at solving the problem of index sequences including interval numbers and fuzzy numbers, this paper presents projection method that is based on the optimal sequence, so ranking of alternatives can be expressed by projection values, and thelinear characteristics of projection is proved. This method solves the ranking problems of sequences containing different number types.However, in real problems of economic and military affairs, the hybrid sequences containing precision numbers, interval numbers and fuzzy numbers are common. This paper provides grey relational model, and proves the model's characteristics, such as: whole, norm, closing, symmetry, and keeping-difference-effect, keeping-order-effect. And the grey relational degree quantity model of multiple index decision making is established, the ranking of alternatives is expressed by relational degree between the project and optimal one.As the hybrid multiple index sequence without weights, this paper establishes a warp function, and the weight sequence ft>; computed makes the total warp maximal. An objective weighting method is proposed.The research findings enrich and develop the theories and methods of multiple attribute decision making, and the two weighting methods are presented.