Intelligent Multi-Attribute Evaluation Of Water Security

Water security is one of today's hot issues in the field of water science. It is related to many subjects and various complex problems. As a joint point of water security system analysis and system decision, water security system evaluation is very important in the application of water security system, the result of which is the basis for decision making. Due to the complicity of water security evaluation, and because the comprehensive evaluation based on common model building and system optimization method which is inclined to classic mathematic methodology cannot deal with practical problems in water security evaluation, in this paper, several intelligent methods combined with classic method are applied to water security evaluation and thus intelligent multi-attribute evaluation of water security is put forward.According to evaluation standards and from the point view of decision makers, multi-attribute evaluation of water security is divided into four categories: water security clustering evaluation, water security ranking evaluation, water security decision-making evaluation, and water security conflict analysis. Aiming at the problem existing in the classic clustering evaluation-Principal Components Analysis, two improved method are proposed; After comparing the merits and demerits of many ranking evaluation methods, the merits of variable fuzzy sets theory is pointed out and the selection of critical point of this theory is discussed; Discussing some TOPSIS (technique for order preference by similarity to ideal solution) functions from the point view of value function, two types of four optimization methods are established to evaluate water security system based on the "difference driving" theory; Water security conflict evaluation is proposed and a improved competitive assessment model based on restriction of Nash equilibrium in multi-criteria decision making is adopted to handle it. Weight assignment is a critical problem in multi-attribute evaluation, and in this paper different weight assigning methods are put forward in the above-mentioned four categories of problems. Objective weight is set according to Principal Components Analysis or TOPSIS model based on the "difference driving" theory, while subjective weight is assigned according to Analytical Hierarchy Process, what's more, variable weight and conflict weight is also discussed. To illustrate the efficiency of these methods, several case studies are practiced and the result is quite satisfactory.