Applications Of Set Pair Analysis To Uncertainty Analysis Of Water Resources

Optimal allocation of water resources is an important means to achieve water resources and water users in different regions of the effective and equitable distribution, so as to achieve sustainable use of water resources. Water resources allocation can achieve to the overall regulation through the water cycle and its impact on the watershed's natural and social factors. But due to a number of factors, the change of hydrology and water resources phenomenon (runoff, flooding, etc.) is extremely complex, with great uncertainty. There is a lot of uncertainty in the entire process of optimal allocation of water resources, including system prediction, simulation, evaluation, reasoning, decision-making and regulation and so on. The Set Pair Analysis is a new method to analysis of uncertainty relations that was first proposed by China's scholar Zhao Ke-qin in national system theory and regional planning meetings at Baotou of Inner Mongolia in 1989 years, based on the philosophy of the unity of opposites and the principle of the universal link. Practice has proved that a large number of applications, the uncertainty analysis in hydrology and water resources with Set Pair Analysis had shown a more comprehensive description of the uncertainty, the concept clear and concise principles calculated to facilitate, as well as the results of an intuitive and reliable. In this paper, bases on detail describing the basic concepts and theories of optimal allocation of water resources and its uncertainty, discusses the research progress and application status quo of the uncertainty analysis in hydrology and water resources by Set Pair Analysis, in-depth studies the feasibility, applicability and specific ways to achieve of the uncertainty analysis in hydrology and water resources by Set Pair Analysis, and the major results are summarized as following:(1) Set pair analysis theory provides a new way for the identified uncertain system, the connection number regression model to forecast that according the set pair analysis theory can significantly improve the prediction accuracy of regression model. For the dynamic of the predictor structure, this paper will use nearest neighbor estimate, by calculating the coefficient of variation of each predictor to determine predictive factors are strong or weak in a certain prediction, then select the strong forecasting function factors dynamically, to eliminate the negative effect from the weak forecasting factors, this reflects well the predictor structure is dynamic. And then a nearest neighbor estimate based dynamic connection number regression model, named NNE-DCNR for short, was established.(2) To characterize the internal distribution rules of annual runoff sequence effectively, and make a more rational division interval, apply the sequence clustering method to establish the wetness-dryness classification criteria of annual runoff sequence. For the characteristics of annual runoff sequence such as the existence of dependencies, randomness and uncertainty, set pair weight based on set pair analysis is proposed. This contains more inherent uncertainty information of annual runoff sequence than the existing self-correlation coefficient weight, express the importance of the lag time step length of Markov chain more reasonably. And then a set pair weight markov chain model based on sequence clustering method for dynamically predicting annual runoff was established.(3) For the fuzzy and uncertainty of interval multi-attribute decision problem, using the binary connection number theory to intuitively describe the relationship between certainty and uncertainty of decision problem. Propose the definition of binary connection number deviation degree, and calculate the total deviation of all the decision-making properties about each decision-making program, then by constructing a single-objective optimization problem to determine the optimal weight vector of each decision attribute. Calculate the deviation degree between decision attribute values and the best program of these attribute values, then can obtain the integrated and weighted deviation degree between the decision-making programs and the best program of these decision-making programs. And then can realize the sorts of pros and cons for each decision-making program. According to this, a binary connection number based interval multi-attribute decision model of water resources was established.(4) The study results show that deeply researching the Set Pair Analysis method in uncertainty analysis of hydrology and water resources, can be further excavated and comprehensively describe the complex structure of hydrology and water resources about uncertainty, but also helped to enrich, improve and develop the theory of hydrology and water resources uncertainty analysis. It is the new developed direction of uncertainty analysis in hydrology and water resources, has a broad application prospects.