New Methods And Applications To Analysis And Calculation Of Water Environment And Water Resource

Analyzing and simulating the development and evolutionary process of water environment and water resource, estimating the degree of pollution, systematically analyzing and calculating the water environment and water resource, are the topics which need to be solved to improve the quality of water environment and the utilizable ability of water resource, as well as can provide science foundation for water environment and water resource management and decision-making. In the present, water environment and water resource analysis and calculation is still in developing, which is difficult in environmental science study. In order to simulate the law of water environment and water resource, and reflect the reality of water quality more objectively and more exactly, it is necessary for science development to use new theories and methods to study water environment and water resource.Under the support of the National Key Project for Basic Research of China (No. 2002CB412301) and the National Natural Science Foundation of China (No. 40271024), this thesis systematically introduces two modern new theories—Projection Pursuit, Support Vector Machine—to systematically study analysis and calculation of water environment and water resource directly, and three optimizing algorithm—Taboo Search, Simulated Annealing, Quantum Genetic Algorithm—to optimize the parameters of new methods of analysis and calculation, with the coupling way of five new methods. The results, performance and characteristics were contrasted in the coupling models. Two improved way of TS was bring forwarded. Through simulating and studying the development of water environment and water resource, all new methods and coupling models were applied to quality assessment and prediction for main river, lake, underground-water and water resource utilizable ability. Results show the practicability of the new methods. Principal findings are concluded as follows:(1) Taboo Search was applied in optimizing the parameters of models of water environment and water resource. Improved Immune Taboo Search and Hybrid Taboo Search was introduced here, ITS was used in predicting regional water resource ability and optimizing the harmony index formula of lake eutrophication; HTS was used in evaluating the level of lake eutrophication, underground water, and in optimizing the simulating formula between water level and water flux.(2) Simulated Annealing was applied in optimizing the parameters of models of water environment and water resource, including index formula of summed power function for evaluating reproducible ability and utilizing ability of water resource, growth exponential formula by S-type of underground water suited to Taiyuan.(3) The application of Quantum Genetic Algorithm was studied in optimizing the parameters of models of water environment and water resource, including water assessment for main cities of Changjiang River, main water basin of Wuhan, and underground water of Taiyuan. Examples show the multiplicity and convergence acceleration of QGA.(4) Projection Pursuit Classification and Projection Pursuit Regression were applied to analysis and calculation of water-environment. PPC was used in water quality assessment of 13 dots of main water area of Wuhan. Parameters of PPR model were optimized by SMART usually, which was complex in calculating. Therefore, a new PPR model with matrix parameters optimized by Immune Evolution Algorithm was put forward and used to predict eutrophic quality trends of Taihu Lake in future 30 years.(5) Parameters of Projection Pursuit model were optimized by TS, SA and QGA respectively. After applying the optimized models in evaluating 13 dots of main water area of Wuhan and water resource utilizable ability of Huaihe basin in China, the evaluation results and the model characteristics were analyzed and contracted.(6) The application of Support Vector Machine was studied in evaluating the water quality of main cities of Changjiang River after illustrating the principle and main characteristics of it.(7) How to choose parameters in SVM is a key technology to improve the precision and complexity, therefore, main parameters of SVM were optimized by TS, and applied to eutrophic quality prediction of Taihu Lake first. Then, in order to grope for parameter selecting technology, algorithm TS, SA and QGA were used to optimize parameters of SVM. Optimized SVM was used in eutrophication grade assessment of 30 lakes in China and 13 dots of main water area of Wuhan.