淮河蚌埠段水环境非线性特征分析与水质预测研究
|
摘要
水环境管理通常包括水质管理和水资源量管理。水环境系统是一个复杂的动态系统，与外界不断进行物质和能量的交换，它具有复杂性属性。对于水环境系统，人们对水环境的内部机制和外部环境的影响和作用机理以及内部和外部的耦合机制以及相互作用尚不清楚的情况下，运用传统的研究方法对水环境管理开展研究显得十分有限。淮河是一个开放的非线性系统，对其水环境非线性特征进行研究，并选择合适的水质预测方法，具有一定的理论和实践意义。
本文在分析河流水环境系统非线性特征的基础上，以淮河蚌埠段为例，研究了其水质的单因素与多因素预测。
（1）河流主要污染因子提取。基于粗糙集理论思想，利用不确定性的近似分类，分析淮河水质数据及其分类关系，客观地确定出淮河干流蚌埠段的主要污染因子及其造成的污染河流的贡献率。
（2）对河流的主要污染因子变化趋势和突变特征分析。淮河污染物浓度有年际变化规律，突变部分是重要信息，往往是严重污染的状态点。小波分析确定出淮河干流蚌埠段主要的污染物年际变化规律和突变特征，提供了一个进一步污染预测和控制的重要依据。
（3）单因素污染因子预测。非线性时间序列预测和轨迹预测在小样本数据的情况下也很有效。分形理论用于单因子污染因子的水质模拟。分形插值函数插值于历史数据，保存演变过程中的一些确定性的东西，由迭代生成过程得出在确定性基础之上的充分的随机行为。实际应用表明，变维分形理论用于水质预测是可行的，且计算速度快，精度高，无收敛性问题，其缺点是不能对多因素并行预测。
（4）多因素污染因子预测。淮河流域水污染往往是一个多因素污染系统，污染因子间有一定的协同和拮抗作用。本文用WNN模型来进行水质多因素预测。小波神经网络结合小波变换的时频局部性质及神经网络的功能，因而具有较强的逼近、容错能力。以河流枯水期水质数据为例进行了水质预测，结果较为理想。
（5）改进的多因素污染因子预测。对于复杂的河流污染系统，本文使用改进的QGA-BP模型，综合利用量子并行计算、遗传算法较好的全局收敛性以及BP神经网络的大规模自适应并行处理、快速学习能力，使量子遗传算法优化的BP神经网络比传统的算法具有更强的并行处理能力和更快的收敛速度。仿真结果表明，改进的量子遗传算法比其他的BP算法预测效率更好。
本文通过对非线性算法的分析、融合和改进，应用多种算法开展水环境非线性特征分析和水质预测研究，对于解决淮河蚌埠段水环境系统评价、分类、预测和决策等实际问题具有重要的理论和现实意义。
Water environment management includes water quality management and waterresources management. Water environment system is a complex dynamic system exchangingmaterials and energy with the outside world continuously. It has the attributes of the complexity. Theinternal mechanism and external environment effects and mechanism of action of and between theinterior and exterior of the interaction and coupling mechanism is not clear. The use of traditionalmethods is very limited to study water environment management, Huaihe is an open nonlinearsystem, it has a certain theoretical and practical significance to study the nonlinear characteristics ofthe water environment and select the appropriate methods of water quality forecast.
Based on the analysis of nonlinear characteristics of river water environmental system, takingHuaihe in Bengbu as an example, the study on single factor and multiple factors prediction of theriver water quality is carried out.
（1）The main pollution factors of river are extracted. Based on the rough set theory, theapproximation classified method for uncertain problems is used. Through the analysis of the waterquality data and its classification relationship of the monitoring points in the Huaihe River, the mainpollution factors and their contribution rate of causing the Huaihe river pollution are determined.
（2）Analysis on river pollutant changing trends and mutation characteristics. River pollutantconcentration of time sequence has certain annual change law, river pollutant concentrationmutations of time sequence are often more important part of information, often are the state points ofserious pollution. It is of great significance to analysis mutation characteristics of river pollutantconcentration of time sequence. Example with Bengbu paragraph in Huaihe River, the river pollutantchanging trends and mutation characteristics are analyzed by using wavelet analysis for riverpollutant concentration of time series. It has very important significance to further forecast andcontrol river pollutant.
（3）Prediction of the single pollution factor. The nonlinear time series forecasting and trackprediction are considered very effective even if using small amount of data. Some certainty things ofevolution process are saved with fractal interpolation functions interpolating in historical records. Itbrings some convenience to water prediction. From the prediction results, it is feasible to usevariable dimension fractal theory for water quality prediction, and the calculating speed andprecision are high.
（4）Prediction of the multi pollution factors. River water quality pollution is oftenmultifactorial pollution system, there are synergistic and antagonistic effects between the pollution factors. In this paper, WNN model is used to simulate the multifactorial water pollution system andpredict multifactorial water quality. WNN is a new neural network model constructed with wavelettheory, which combines the time-frequency characters of wavelet transform and the function ofneural network, thus it has the strong approximation and fault tolerance capability. WNN has betterproperties than the traditional neural network.
（5）Improved prediction of the multi pollution factors. In this paper, the improved QGA-BPmodel is used in the complex river pollution system, with comprehensive utilization of the betterglobal convergence properties of quantum computation, quantum entanglement and geneticalgorithm, and with comprehensive utilization of the adaptive large-scale parallel processing and fastlearning ability of the BP neural network, so that the BP neural network optimized with the quantumgenetic algorithm is better than the traditional algorithm, it has stronger parallel processing abilityand faster convergence speed. The simulation results show that the improved QGA-BP algorithm isbetter than other BP algorithm, the improved algorithm has better prediction efficiency, and theoperation is not divergent.
In this article, series of algorithm improvement and integration strategy are studied withnonlinear algorithm analysis and fusion, in order to find efficient and quick solution. A variety ofalgorithm is applied for higher-dimensional function optimization and optimizing performanceanalysis of complex system model. It has important theoretical and practical significance to solveevaluation, classification, prediction and decision-making for Huaihe river environmental system inBengbu.
本文在分析河流水环境系统非线性特征的基础上，以淮河蚌埠段为例，研究了其水质的单因素与多因素预测。
（1）河流主要污染因子提取。基于粗糙集理论思想，利用不确定性的近似分类，分析淮河水质数据及其分类关系，客观地确定出淮河干流蚌埠段的主要污染因子及其造成的污染河流的贡献率。
（2）对河流的主要污染因子变化趋势和突变特征分析。淮河污染物浓度有年际变化规律，突变部分是重要信息，往往是严重污染的状态点。小波分析确定出淮河干流蚌埠段主要的污染物年际变化规律和突变特征，提供了一个进一步污染预测和控制的重要依据。
（3）单因素污染因子预测。非线性时间序列预测和轨迹预测在小样本数据的情况下也很有效。分形理论用于单因子污染因子的水质模拟。分形插值函数插值于历史数据，保存演变过程中的一些确定性的东西，由迭代生成过程得出在确定性基础之上的充分的随机行为。实际应用表明，变维分形理论用于水质预测是可行的，且计算速度快，精度高，无收敛性问题，其缺点是不能对多因素并行预测。
（4）多因素污染因子预测。淮河流域水污染往往是一个多因素污染系统，污染因子间有一定的协同和拮抗作用。本文用WNN模型来进行水质多因素预测。小波神经网络结合小波变换的时频局部性质及神经网络的功能，因而具有较强的逼近、容错能力。以河流枯水期水质数据为例进行了水质预测，结果较为理想。
（5）改进的多因素污染因子预测。对于复杂的河流污染系统，本文使用改进的QGA-BP模型，综合利用量子并行计算、遗传算法较好的全局收敛性以及BP神经网络的大规模自适应并行处理、快速学习能力，使量子遗传算法优化的BP神经网络比传统的算法具有更强的并行处理能力和更快的收敛速度。仿真结果表明，改进的量子遗传算法比其他的BP算法预测效率更好。
本文通过对非线性算法的分析、融合和改进，应用多种算法开展水环境非线性特征分析和水质预测研究，对于解决淮河蚌埠段水环境系统评价、分类、预测和决策等实际问题具有重要的理论和现实意义。
Water environment management includes water quality management and waterresources management. Water environment system is a complex dynamic system exchangingmaterials and energy with the outside world continuously. It has the attributes of the complexity. Theinternal mechanism and external environment effects and mechanism of action of and between theinterior and exterior of the interaction and coupling mechanism is not clear. The use of traditionalmethods is very limited to study water environment management, Huaihe is an open nonlinearsystem, it has a certain theoretical and practical significance to study the nonlinear characteristics ofthe water environment and select the appropriate methods of water quality forecast.
Based on the analysis of nonlinear characteristics of river water environmental system, takingHuaihe in Bengbu as an example, the study on single factor and multiple factors prediction of theriver water quality is carried out.
（1）The main pollution factors of river are extracted. Based on the rough set theory, theapproximation classified method for uncertain problems is used. Through the analysis of the waterquality data and its classification relationship of the monitoring points in the Huaihe River, the mainpollution factors and their contribution rate of causing the Huaihe river pollution are determined.
（2）Analysis on river pollutant changing trends and mutation characteristics. River pollutantconcentration of time sequence has certain annual change law, river pollutant concentrationmutations of time sequence are often more important part of information, often are the state points ofserious pollution. It is of great significance to analysis mutation characteristics of river pollutantconcentration of time sequence. Example with Bengbu paragraph in Huaihe River, the river pollutantchanging trends and mutation characteristics are analyzed by using wavelet analysis for riverpollutant concentration of time series. It has very important significance to further forecast andcontrol river pollutant.
（3）Prediction of the single pollution factor. The nonlinear time series forecasting and trackprediction are considered very effective even if using small amount of data. Some certainty things ofevolution process are saved with fractal interpolation functions interpolating in historical records. Itbrings some convenience to water prediction. From the prediction results, it is feasible to usevariable dimension fractal theory for water quality prediction, and the calculating speed andprecision are high.
（4）Prediction of the multi pollution factors. River water quality pollution is oftenmultifactorial pollution system, there are synergistic and antagonistic effects between the pollution factors. In this paper, WNN model is used to simulate the multifactorial water pollution system andpredict multifactorial water quality. WNN is a new neural network model constructed with wavelettheory, which combines the time-frequency characters of wavelet transform and the function ofneural network, thus it has the strong approximation and fault tolerance capability. WNN has betterproperties than the traditional neural network.
（5）Improved prediction of the multi pollution factors. In this paper, the improved QGA-BPmodel is used in the complex river pollution system, with comprehensive utilization of the betterglobal convergence properties of quantum computation, quantum entanglement and geneticalgorithm, and with comprehensive utilization of the adaptive large-scale parallel processing and fastlearning ability of the BP neural network, so that the BP neural network optimized with the quantumgenetic algorithm is better than the traditional algorithm, it has stronger parallel processing abilityand faster convergence speed. The simulation results show that the improved QGA-BP algorithm isbetter than other BP algorithm, the improved algorithm has better prediction efficiency, and theoperation is not divergent.
In this article, series of algorithm improvement and integration strategy are studied withnonlinear algorithm analysis and fusion, in order to find efficient and quick solution. A variety ofalgorithm is applied for higher-dimensional function optimization and optimizing performanceanalysis of complex system model. It has important theoretical and practical significance to solveevaluation, classification, prediction and decision-making for Huaihe river environmental system inBengbu.
引文
[1]2005年中国环境状况公报:淡水环境.国家环保总局网站.
[2]淮河流域水污染成因及防范对策.中国水利,2005,22.
[3]水利部淮河水利委员会.2004年淮河流域城镇入河排污量监测报告[R].2005.
[4]李春.国外地下水资源的保护与管理研究动态.中国人口·资源与环境,2001年第11卷.
[5] USNRC. Innovations in ground water and soil cleanup from concept to commercialization.National Academy Press,1997.
[6] USNRC. Building an Effective Environmental Management Science Program: FinalAssessment. National Academy Press,1997.
[7]刘清. Rough集及Rough推理[M].北京:科学出版社,2001:11-38.
[8]张文修,仇国芳.基于粗糙集的不确定决策[M].北京:清华大学出版社,2005.
[9]周伟,桂林,周林. MATLAB小波分析高级技术[M].西安:西安电子科技大学出版社,2006:51-63.
[10]蒋晓辉,刘昌明.基于小波分析的径向基神经网络年径流量预测[J].应用科学学报,2004,22(3):411-414.
[11]刘明才.小波分析及其应用.北京:清华大学出版社,2005.
[12] Dwight F. Mix, Kraig J. Olejniczak(杨志华,杨力华译).小波基础及应用教程.北京:机械工业出版社,2006.
[13]张济忠.分形[M].北京:清华大学出版社,1995.
[14] Paul Meakin.分形、标度及远离平衡态的增长:[英文影印][M].北京:清华大学出版社,2000.
[15] Kenneth Falconer(曾文曲译).分形几何:数学基础及其应用.北京:人民邮电出版社,2007.
[16]袁昊,刘廷建,邓睿.基于粗糙集数据挖掘技术开发的用水量计划系统[J].科技情报开发与经济,2004(2).
[17]谢敦礼,李浩,叶福根,等.数据挖掘在城市供水系统中的一个应用[J].系统工程理论与实践,2003(6).
[18]刘廷建,王自豪,郑朝霞.基于关联规则的数据挖掘在供水管网故障诊断中的运用[J].科技情报开发与经济,2005(14).
[19]党建武.神经网络技术及应用[M].北京:中国铁道出版社,2000.
[20]崔宝侠,段勇,高鸿雁,等.神经网络在水质模型中的应用[J].沈阳工业大学学报,2003,25(3):220-222.
[21] Thomas Dean, James Allen, Yiannis Aloimonos(顾国昌,等译).人工智能:理论与实践.北京:电子工业出版社,2004.
[22] M. Tim Jones(黄厚宽,等译).人工智能:a systems approach.北京:电子工业出版社,2010.
[23]蔡自兴,徐光祐.人工智能及其应用:principles and applications(第3版).北京:清华大学出版社,2004.
[24]宋华兵,张新政.内嵌神经网络的不确定水质模型研究[J].水资源与水工程学报,2008,19(1):12-14.
[25]蒋霖,李向民,周海峰.多小波神经网络在变形预测中的应用[J].山西建筑,2008,34(1):13-14.
[26]叶春明,马慧民,李丹,等. BP神经网络在供应链绩效指标评价中的应用研究[J].工业工程与管理,2005(5):35-38.
[27]丁晶,邓育仁.随机水文学[M].成都:成都科技大学出版社,1988.
[28]钟锦.基于演化博弈的淮河流域水环境管理研究[D].合肥:合肥工业大学,2008.
[29]姚重华.环境工程仿真与控制[M].北京:高等教育出版社,2001.
[30]张文修等.粗糙集理论与方法[M].科学出版社,2001.
[31]苗夺谦,李道国.粗糙集理论、算法与应用[M].北京:清华大学出版社,2008.
[32]张利,卢秀颖,吴华玉,郝胜智.基于粗糙集的启发式值约简的改进算法[J].仪器仪表学报,2009,(01).
[33]高晓康,薛军,施雨辰.一种基于粗糙集理论的决策规则提取方法[J].上海应用技术学院学报(自然科学版),2009,(03).
[34]袁智敏,黄庆,汪江洪.一种新的综合评价方法——粗糙集灰色聚类评价[J].统计与决策,2005,(17).
[35]吴海涛,魏长宝.决策系统属性约简优化算法研究[J].计算机应用与软件,2009,(07).
[36]叶东毅,陈昭炯.不相容决策表全部属性约简计算的一个改进方法[J].小型微型计算机系统,2006,(10).
[37]谭天乐,宋执环,李平.基于粗糙集的故障诊断方法[J].浙江大学学报(工学版),2003,(01).
[38]李建平,唐远炎.小波分析方法的应用[M].重庆:重庆大学出版社,1999.
[39]赵松年,熊小芸.小波变换与小波分析[M].北京:电子工业出版社,1996.
[40] J. Morlet, G. Arens, E. Forgeau, et al, Wave propagation and sampling theory-part2: samplingtheory and complex waves. Geophysics,1982,47(2):222-236.
[41] A. Grossmann, and J. Morlet, Decomposition of Hardy functions into square integrablewavelets of constant shape, SIMA. J. Math. Anal,1984,15(4):723-726.
[42]葛哲学,沙威编著小波分析理论与MATLAB R2007实现北京:电子工业出版社,2007
[43] S. G. Mallat, Multiresolution representation and wavelets,[Ph.D.Thesis], Philadelphia:University of Pennsylvania,1988.
[44] S. G. Mallat, A theory for multiresolution signal decomposition: the wavelet representation,IEEE Trans. on Pater Analysis and Machine Intelligence,1989,11(7):674-693.
[45] S. G. Mallat, Multifrequency channel decompositions of images and wavelet models, IEEEtrans. on Acoustics. Speech and Signal Processing,1989,37(12):2091-2110.
[46] I. Daubechies,Ten Lectures on Wavelet, Capital City Press,1992.
[47] Coiftnan R., Wickerhauser M., Entropy-based algorithms for best basis selection, IEEE Trans.on Information Theory,1992,38(3):713-718.
[48] C. Taswell, Handbook of Wavelet Transform Algorithms, Brikhauser, Boston,1996.
[49] M.Frisch and H. Messer, The Use of Wavelet Transform in the Detection of an UnknownTransient Signals, IEEE Trams. On Information Theory,38(2),1992.
[50]李建平.小波分析的一些有前景的应用领域[J].重庆大学学报(自然科学版),1999,22(1):121-125.
[51]李建平.小波分析与信号处理——理论、应用及软件实现[M].重庆:重庆大学出版社,1997.
[52]薛小杰,蒋晓辉,黄强,等.小波分析在水文序列趋势分析中的应用[J].应用科学学报,2002,(4).
[53]胡昌华,张军波,夏军.基于MATLAB的系统分析与设计——小波分析[M].西安:西安电子科技大学出版社,1999.
[54]王文圣,丁晶,李跃清.水文小波分析[J].北京:化学工业出版社.
[55]王文圣,赵太想,丁晶.基于连续小波变换的水文序列变化特征研究[J].四川大学学报,2004,(4).
[56]王文圣,丁晶,向红莲.水文水资源时间序列多时间尺度分析的小波变换[J].四川大学学报(工程科学版),2002,.
[57]郑昱,张闻胜.基于小波变换的水文序列的近似周期检测法[J].水文,1999,(6).
[58]刘素一,廖家平.径流预测中的非线性分析方法[J].湖北工学院学报,2003,(1).
[59]陈柳.小波分析和神经网络应用于大气污染预测的研究[D].2006.
[60]崔锦泰.小波分析导论[M].西安:西安交通大学出版社,1995.
[61]李贤彬,丁晶,李后强.水文时间序列的子波分析法[J].水科学进展,1999,10(2):144-149.
[62]李贤彬,丁晶,李后强.水文序列Hurst系数的子波估计法[J].水利学报,2000(2):6-9.
[63]彭玉华.小波变换与工程应用[M].北京:科学出版社,1999.
[64]邓自旺,林振山,周晓兰.西安市近50年来气候变化多时间尺度分析[J].高原气象,1997,16(1):81-93.
[65]覃光华,丁晶,刘国栋.自适应BP算法及其在河道洪水预报上的应用[J].水科学进展,2002,(2).
[66]王文圣,李跃清,向红莲.基于小波分析的组合随机模型及其在径流预测中的应用[J].高原气象,2004,23(S):146-149.
[67]胡平,康玲.一种非线性组合预测方法在径流预测中的应用[J].中国农村水利水电,2004(3):38-40.
[68]王文圣,袁鹏,丁晶.小波分析及其在日流量过程随机模拟中的应用[J].水利学报,2000(11):43-48.
[69]钱学森.再谈开放的复杂巨系统.模式识别与人工智能,1991,4(1):1-4.
[70]沙震,阮火军.分形与拟合[M].浙江:浙江大学出版社,2005.
[71] Barnsley M F. Fractal functions and interpolation[J]. Const. Approx,1986(2):303-329.
[72]陶涛,刘遂庆.基于分形理论的需水量预测方法[J].同济大学学报(自然科学版),2004,32(12):1647-1650.
[73]付显华.平移分形分析和预测月平均海面水温[J].海洋预报,1996,13(2):63-68.
[74]冯利华.基于R/S分析的水资源预测[J].系统工程理论践,2002,4:115-118.
[75]赵健,雷蕾,蒲小勤.分形理论及其在信号处理中的应用[M].北京:清华大学出版社,2008.
[76]杨位钦,顾岚.时间序列分析和动态数据建模[M].北京:北京工业学院出版社,1986.
[77]申维.分形混沌与矿产预测[M].北京:地质出版社,2002.
[78]付显华,付安捷.用分形方法预测石油股票价格和指数[J].中国海洋平台,2002,17(6):41-45.
[79]陈晓,朱耀庭,朱光喜.基于预测模型的分形图象压缩编码方法[J].中国图象图形学报,1998,3(6):471-475.
[80] A.Scotti, C.Meneveau. A fractal model for large eddy simulation of turbulent flow[J]. PhysicaD,1999,127:198-232.
[81] L.Dalla, V. Drakopoulos. On the parameter identification problem in the plane and polar fractalinterpolation functions[J]. Journal of approximation theory,1999,101:289-302.
[82] R.Pitchumani, B.Ramakrishnan. A fractal geometry model evaluating permeabilities of porousperforms used in liquid composite molding[J]. International journal of heat and masstransfer,1999,42:2219-2232.
[83] Kazuyoshi Nanjo, Hiroyuki Nagahama. Fractal properties of spatial distributions of aftershocksand active faults [J].Chaos, solitons and fractals,2004,19:387-397
[84]冯志刚,王磊.一维分形插值函数的小波类型级数表示及误差估计[J].江苏大学学报(自然科学版),2005(4).
[85]谢和平,冯志刚.岩石断面的分形插值研究.第八届全国现代数学和力学学术会议,2001.
[86]李江.城市空间形态的分形维数及应用[J].武汉大学学报:工学版,2005,38(3):99-104.
[87]郑长青.研究城市快速公路交通流的分形数学方法[J].交通运输系统工程与信息,2006,6(1):96-99.
[88]宋春林,冯端,刘富强.变步长和变阈值的分形小波图像压缩算法.2007(14).
[89] MANDELBROT B How long is the coast of the Britain? Statistical self-similarity andfractional dimension1967(375)
[90]郝柏林著混沌与分形：郝柏林科普文集上海科学技术出版公司,2004
[91] Simon Haykin.神经网络原理[M].北京:机械工业出版社,2004.
[92]张立明.人工神经网络的模型及其应用[M].上海:复旦大学出版社,1992.
[93]常沁春. BP神经网络改进算法预测水质浓度的应用[J]甘肃环境研究与监测,2002,15(3):186-188,226.
[94]崔宝侠,段勇,高鸿雁,等.神经网络在水质模型中的应用[J].沈阳工业大学学报,2003,25(3):220-222.
[95]陈雄波,唐洪武.用BP神经网络原理对水流挟沙力的研究[J].泥沙研究,2004,1:29-34.
[96]杨延竹,赵学增,王伟杰,等.基于模糊神经网络的彩色图像滤波方法[J].控制与决策,2004,19(1):69-72.
[97]丛爽.面向MATLAB工具箱的神经网络理论与应用[M].合肥:中国科学技术大学出版社,2003.
[98]胡坚,严敏,刘俊萍,等.径向基函数人工神经网络预测污水处理厂出水水质[J].浙江工业大学学报,2006,34(6):633-636.
[99]牛东晓,邢棉.时间序列的小波神经网络预测模型的研究[J].系统工程理论与实践,1999,(5):89-92.
[100] J. Morlet, G. Arens, E. Forgeau, et al, Wave propagation and sampling theory-part1: complezsignal and scatering in multilayered media, Geophysics,1982,47(2):203-221.
[101] S. G. Mallat, A theory for multiresolution signal decomposition: the wavelet representation,IEEE Trans. on Patern Analysis and Machine Intelligence,1989,11(7):674-693.
[102] Marina Campolo, et al.Forecasting river flow rate during low-flow period using neuralnetwork [J]. Water resource research,1999,35(11):3547-3552.
[103] T.R.Neelakantan, N.VPundarikanthan.Neural network-based simulation-optimization modelfor reservoir operation[J]. Journal of water resources planning and management,2000,126(2):57-64.
[104] Sharad Kumar Jain. Development of integrated sediment rating curves using ANNs[J].Journal of hydraulic engineering,2001,127(3):181-193.
[105]董长虹. MATLAB小波分析工具箱原理与应用[M].国防工业出版社.
[106] MATLAB6.5辅助神经网络分析与设计[M].北京:电子工业出版社.
[107] MATLAB在环境科学中的应用[M].北京:化学工业出版社,2004.
[108]应海松.小波神经网络在铁矿石检验中的应用[M].冶金工业出版社.
[109]李贤彬,丁晶,李后强.基于小波变换序列的神经网络组合预测[J].水利学报,1999(2):1-4.
[110] Delyon B, Juditsky A, Benveniste A. Accuracy analysis for wavelet approximations [J]. IEEETrans on NN,1995,6(2):332-348.
[111] Fraser A M. Independent Coordinates for Strange Attractors from Mutual Information [J].Physical Review A,1986,32(2):1134-1140.
[112]樊启斌.小波变换及其在遥感图像压缩中的应用[D].武汉:武汉大学,2001
[113] George W R G. Nonlinear decision weights in choice under uncer2tainty [J]. ManagementScience,1999,45:1.
[114]高成. MATLAB小波分析与应用[M].北京:国防工业出版社,2007.
[115]飞思科技产品研发中心. MATLAB6.5辅助小波分析与应用[M].北京:电子工业出版社,2003.
[116] Cybenko G. Approximation by superpositions of a sigmoid function[J]. Math Control,Signal and Systems,1989,2(4):201-205.
[117] ZHANGQing2hua, Benveniste Albert. Wavelet Networks[J]. IEEE Trans on NN,1992,3(6):311-31
[118]秦前清,杨宗凯.实用小波分析[M].西安:西安电子科技大学出版社,1995.
[119]崔锦泰.程正兴译.小波分析导论[M].西安:西安交通大学出版社,1995.
[120] Bakshi B R, Stephanopoulos G.Wavenet: a multiresolution hierachical neural network withlocalized learning [J]. AICHEJ,1993,39(1):57-81.
[121] Daubechies I.Orthonormal base of compactly supported wavelets [J]. Communications onPure and Applied Mathematics,1988,41(2):909-996.
[122]武君.河流水质模拟预测的常用方法研究与新方法探索——以淮河安徽段为例[D].合肥:合肥工业大学,2005.
[123]钟锦.基于演化博弈的淮河流域水环境管理研究[D].合肥:合肥工业大学,2008.
[124]党建武.神经网络技术及应用[M].北京:中国铁道出版社,2000.
[125]丛爽.面向MATLAB工具箱的神经网络理论与应用[M].合肥:中国科学技术大学出版社,2003.
[126] Tony H. Quantum computing: an introduction[J]. Computing&Control Engineering Journal,1996;10(3):105-112.
[127]雷英杰,张善文,李续武,等. Matlab遗传算法工具箱及应用[M].西安:西安电子科技大学出版社,2005.
[128]周明,孙树栋.遗传算法原理及应用[M].国防工业出版社,2002.
[129]周正武,丁同梅,田毅红,等. Matlab遗传算法优化工具箱(GAOT)的研究与应用[J].机械研究与应用,2006,6(19):69-71.
[130] Tony H. Quantum computing: an introduction[J]. Computing&Control Engineering Journal,1996;10(3):105-112.
[131] Han K H, Park K H, Lee C H,et al. Parallel quantum-inspired genetic algorithm forcombinatorial optimization problems[A]. Proceedings of IEEE International Conference onEvolutionary Computation[C]. Piscataway: IEEE Press,2001:1442-1429.
[132]李士勇,李盼池.量子计算与量子优化算法[M].哈尔滨:哈尔滨工业大学出版社,2009.
[2]淮河流域水污染成因及防范对策.中国水利,2005,22.
[3]水利部淮河水利委员会.2004年淮河流域城镇入河排污量监测报告[R].2005.
[4]李春.国外地下水资源的保护与管理研究动态.中国人口·资源与环境,2001年第11卷.
[5] USNRC. Innovations in ground water and soil cleanup from concept to commercialization.National Academy Press,1997.
[6] USNRC. Building an Effective Environmental Management Science Program: FinalAssessment. National Academy Press,1997.
[7]刘清. Rough集及Rough推理[M].北京:科学出版社,2001:11-38.
[8]张文修,仇国芳.基于粗糙集的不确定决策[M].北京:清华大学出版社,2005.
[9]周伟,桂林,周林. MATLAB小波分析高级技术[M].西安:西安电子科技大学出版社,2006:51-63.
[10]蒋晓辉,刘昌明.基于小波分析的径向基神经网络年径流量预测[J].应用科学学报,2004,22(3):411-414.
[11]刘明才.小波分析及其应用.北京:清华大学出版社,2005.
[12] Dwight F. Mix, Kraig J. Olejniczak(杨志华,杨力华译).小波基础及应用教程.北京:机械工业出版社,2006.
[13]张济忠.分形[M].北京:清华大学出版社,1995.
[14] Paul Meakin.分形、标度及远离平衡态的增长:[英文影印][M].北京:清华大学出版社,2000.
[15] Kenneth Falconer(曾文曲译).分形几何:数学基础及其应用.北京:人民邮电出版社,2007.
[16]袁昊,刘廷建,邓睿.基于粗糙集数据挖掘技术开发的用水量计划系统[J].科技情报开发与经济,2004(2).
[17]谢敦礼,李浩,叶福根,等.数据挖掘在城市供水系统中的一个应用[J].系统工程理论与实践,2003(6).
[18]刘廷建,王自豪,郑朝霞.基于关联规则的数据挖掘在供水管网故障诊断中的运用[J].科技情报开发与经济,2005(14).
[19]党建武.神经网络技术及应用[M].北京:中国铁道出版社,2000.
[20]崔宝侠,段勇,高鸿雁,等.神经网络在水质模型中的应用[J].沈阳工业大学学报,2003,25(3):220-222.
[21] Thomas Dean, James Allen, Yiannis Aloimonos(顾国昌,等译).人工智能:理论与实践.北京:电子工业出版社,2004.
[22] M. Tim Jones(黄厚宽,等译).人工智能:a systems approach.北京:电子工业出版社,2010.
[23]蔡自兴,徐光祐.人工智能及其应用:principles and applications(第3版).北京:清华大学出版社,2004.
[24]宋华兵,张新政.内嵌神经网络的不确定水质模型研究[J].水资源与水工程学报,2008,19(1):12-14.
[25]蒋霖,李向民,周海峰.多小波神经网络在变形预测中的应用[J].山西建筑,2008,34(1):13-14.
[26]叶春明,马慧民,李丹,等. BP神经网络在供应链绩效指标评价中的应用研究[J].工业工程与管理,2005(5):35-38.
[27]丁晶,邓育仁.随机水文学[M].成都:成都科技大学出版社,1988.
[28]钟锦.基于演化博弈的淮河流域水环境管理研究[D].合肥:合肥工业大学,2008.
[29]姚重华.环境工程仿真与控制[M].北京:高等教育出版社,2001.
[30]张文修等.粗糙集理论与方法[M].科学出版社,2001.
[31]苗夺谦,李道国.粗糙集理论、算法与应用[M].北京:清华大学出版社,2008.
[32]张利,卢秀颖,吴华玉,郝胜智.基于粗糙集的启发式值约简的改进算法[J].仪器仪表学报,2009,(01).
[33]高晓康,薛军,施雨辰.一种基于粗糙集理论的决策规则提取方法[J].上海应用技术学院学报(自然科学版),2009,(03).
[34]袁智敏,黄庆,汪江洪.一种新的综合评价方法——粗糙集灰色聚类评价[J].统计与决策,2005,(17).
[35]吴海涛,魏长宝.决策系统属性约简优化算法研究[J].计算机应用与软件,2009,(07).
[36]叶东毅,陈昭炯.不相容决策表全部属性约简计算的一个改进方法[J].小型微型计算机系统,2006,(10).
[37]谭天乐,宋执环,李平.基于粗糙集的故障诊断方法[J].浙江大学学报(工学版),2003,(01).
[38]李建平,唐远炎.小波分析方法的应用[M].重庆:重庆大学出版社,1999.
[39]赵松年,熊小芸.小波变换与小波分析[M].北京:电子工业出版社,1996.
[40] J. Morlet, G. Arens, E. Forgeau, et al, Wave propagation and sampling theory-part2: samplingtheory and complex waves. Geophysics,1982,47(2):222-236.
[41] A. Grossmann, and J. Morlet, Decomposition of Hardy functions into square integrablewavelets of constant shape, SIMA. J. Math. Anal,1984,15(4):723-726.
[42]葛哲学,沙威编著小波分析理论与MATLAB R2007实现北京:电子工业出版社,2007
[43] S. G. Mallat, Multiresolution representation and wavelets,[Ph.D.Thesis], Philadelphia:University of Pennsylvania,1988.
[44] S. G. Mallat, A theory for multiresolution signal decomposition: the wavelet representation,IEEE Trans. on Pater Analysis and Machine Intelligence,1989,11(7):674-693.
[45] S. G. Mallat, Multifrequency channel decompositions of images and wavelet models, IEEEtrans. on Acoustics. Speech and Signal Processing,1989,37(12):2091-2110.
[46] I. Daubechies,Ten Lectures on Wavelet, Capital City Press,1992.
[47] Coiftnan R., Wickerhauser M., Entropy-based algorithms for best basis selection, IEEE Trans.on Information Theory,1992,38(3):713-718.
[48] C. Taswell, Handbook of Wavelet Transform Algorithms, Brikhauser, Boston,1996.
[49] M.Frisch and H. Messer, The Use of Wavelet Transform in the Detection of an UnknownTransient Signals, IEEE Trams. On Information Theory,38(2),1992.
[50]李建平.小波分析的一些有前景的应用领域[J].重庆大学学报(自然科学版),1999,22(1):121-125.
[51]李建平.小波分析与信号处理——理论、应用及软件实现[M].重庆:重庆大学出版社,1997.
[52]薛小杰,蒋晓辉,黄强,等.小波分析在水文序列趋势分析中的应用[J].应用科学学报,2002,(4).
[53]胡昌华,张军波,夏军.基于MATLAB的系统分析与设计——小波分析[M].西安:西安电子科技大学出版社,1999.
[54]王文圣,丁晶,李跃清.水文小波分析[J].北京:化学工业出版社.
[55]王文圣,赵太想,丁晶.基于连续小波变换的水文序列变化特征研究[J].四川大学学报,2004,(4).
[56]王文圣,丁晶,向红莲.水文水资源时间序列多时间尺度分析的小波变换[J].四川大学学报(工程科学版),2002,.
[57]郑昱,张闻胜.基于小波变换的水文序列的近似周期检测法[J].水文,1999,(6).
[58]刘素一,廖家平.径流预测中的非线性分析方法[J].湖北工学院学报,2003,(1).
[59]陈柳.小波分析和神经网络应用于大气污染预测的研究[D].2006.
[60]崔锦泰.小波分析导论[M].西安:西安交通大学出版社,1995.
[61]李贤彬,丁晶,李后强.水文时间序列的子波分析法[J].水科学进展,1999,10(2):144-149.
[62]李贤彬,丁晶,李后强.水文序列Hurst系数的子波估计法[J].水利学报,2000(2):6-9.
[63]彭玉华.小波变换与工程应用[M].北京:科学出版社,1999.
[64]邓自旺,林振山,周晓兰.西安市近50年来气候变化多时间尺度分析[J].高原气象,1997,16(1):81-93.
[65]覃光华,丁晶,刘国栋.自适应BP算法及其在河道洪水预报上的应用[J].水科学进展,2002,(2).
[66]王文圣,李跃清,向红莲.基于小波分析的组合随机模型及其在径流预测中的应用[J].高原气象,2004,23(S):146-149.
[67]胡平,康玲.一种非线性组合预测方法在径流预测中的应用[J].中国农村水利水电,2004(3):38-40.
[68]王文圣,袁鹏,丁晶.小波分析及其在日流量过程随机模拟中的应用[J].水利学报,2000(11):43-48.
[69]钱学森.再谈开放的复杂巨系统.模式识别与人工智能,1991,4(1):1-4.
[70]沙震,阮火军.分形与拟合[M].浙江:浙江大学出版社,2005.
[71] Barnsley M F. Fractal functions and interpolation[J]. Const. Approx,1986(2):303-329.
[72]陶涛,刘遂庆.基于分形理论的需水量预测方法[J].同济大学学报(自然科学版),2004,32(12):1647-1650.
[73]付显华.平移分形分析和预测月平均海面水温[J].海洋预报,1996,13(2):63-68.
[74]冯利华.基于R/S分析的水资源预测[J].系统工程理论践,2002,4:115-118.
[75]赵健,雷蕾,蒲小勤.分形理论及其在信号处理中的应用[M].北京:清华大学出版社,2008.
[76]杨位钦,顾岚.时间序列分析和动态数据建模[M].北京:北京工业学院出版社,1986.
[77]申维.分形混沌与矿产预测[M].北京:地质出版社,2002.
[78]付显华,付安捷.用分形方法预测石油股票价格和指数[J].中国海洋平台,2002,17(6):41-45.
[79]陈晓,朱耀庭,朱光喜.基于预测模型的分形图象压缩编码方法[J].中国图象图形学报,1998,3(6):471-475.
[80] A.Scotti, C.Meneveau. A fractal model for large eddy simulation of turbulent flow[J]. PhysicaD,1999,127:198-232.
[81] L.Dalla, V. Drakopoulos. On the parameter identification problem in the plane and polar fractalinterpolation functions[J]. Journal of approximation theory,1999,101:289-302.
[82] R.Pitchumani, B.Ramakrishnan. A fractal geometry model evaluating permeabilities of porousperforms used in liquid composite molding[J]. International journal of heat and masstransfer,1999,42:2219-2232.
[83] Kazuyoshi Nanjo, Hiroyuki Nagahama. Fractal properties of spatial distributions of aftershocksand active faults [J].Chaos, solitons and fractals,2004,19:387-397
[84]冯志刚,王磊.一维分形插值函数的小波类型级数表示及误差估计[J].江苏大学学报(自然科学版),2005(4).
[85]谢和平,冯志刚.岩石断面的分形插值研究.第八届全国现代数学和力学学术会议,2001.
[86]李江.城市空间形态的分形维数及应用[J].武汉大学学报:工学版,2005,38(3):99-104.
[87]郑长青.研究城市快速公路交通流的分形数学方法[J].交通运输系统工程与信息,2006,6(1):96-99.
[88]宋春林,冯端,刘富强.变步长和变阈值的分形小波图像压缩算法.2007(14).
[89] MANDELBROT B How long is the coast of the Britain? Statistical self-similarity andfractional dimension1967(375)
[90]郝柏林著混沌与分形：郝柏林科普文集上海科学技术出版公司,2004
[91] Simon Haykin.神经网络原理[M].北京:机械工业出版社,2004.
[92]张立明.人工神经网络的模型及其应用[M].上海:复旦大学出版社,1992.
[93]常沁春. BP神经网络改进算法预测水质浓度的应用[J]甘肃环境研究与监测,2002,15(3):186-188,226.
[94]崔宝侠,段勇,高鸿雁,等.神经网络在水质模型中的应用[J].沈阳工业大学学报,2003,25(3):220-222.
[95]陈雄波,唐洪武.用BP神经网络原理对水流挟沙力的研究[J].泥沙研究,2004,1:29-34.
[96]杨延竹,赵学增,王伟杰,等.基于模糊神经网络的彩色图像滤波方法[J].控制与决策,2004,19(1):69-72.
[97]丛爽.面向MATLAB工具箱的神经网络理论与应用[M].合肥:中国科学技术大学出版社,2003.
[98]胡坚,严敏,刘俊萍,等.径向基函数人工神经网络预测污水处理厂出水水质[J].浙江工业大学学报,2006,34(6):633-636.
[99]牛东晓,邢棉.时间序列的小波神经网络预测模型的研究[J].系统工程理论与实践,1999,(5):89-92.
[100] J. Morlet, G. Arens, E. Forgeau, et al, Wave propagation and sampling theory-part1: complezsignal and scatering in multilayered media, Geophysics,1982,47(2):203-221.
[101] S. G. Mallat, A theory for multiresolution signal decomposition: the wavelet representation,IEEE Trans. on Patern Analysis and Machine Intelligence,1989,11(7):674-693.
[102] Marina Campolo, et al.Forecasting river flow rate during low-flow period using neuralnetwork [J]. Water resource research,1999,35(11):3547-3552.
[103] T.R.Neelakantan, N.VPundarikanthan.Neural network-based simulation-optimization modelfor reservoir operation[J]. Journal of water resources planning and management,2000,126(2):57-64.
[104] Sharad Kumar Jain. Development of integrated sediment rating curves using ANNs[J].Journal of hydraulic engineering,2001,127(3):181-193.
[105]董长虹. MATLAB小波分析工具箱原理与应用[M].国防工业出版社.
[106] MATLAB6.5辅助神经网络分析与设计[M].北京:电子工业出版社.
[107] MATLAB在环境科学中的应用[M].北京:化学工业出版社,2004.
[108]应海松.小波神经网络在铁矿石检验中的应用[M].冶金工业出版社.
[109]李贤彬,丁晶,李后强.基于小波变换序列的神经网络组合预测[J].水利学报,1999(2):1-4.
[110] Delyon B, Juditsky A, Benveniste A. Accuracy analysis for wavelet approximations [J]. IEEETrans on NN,1995,6(2):332-348.
[111] Fraser A M. Independent Coordinates for Strange Attractors from Mutual Information [J].Physical Review A,1986,32(2):1134-1140.
[112]樊启斌.小波变换及其在遥感图像压缩中的应用[D].武汉:武汉大学,2001
[113] George W R G. Nonlinear decision weights in choice under uncer2tainty [J]. ManagementScience,1999,45:1.
[114]高成. MATLAB小波分析与应用[M].北京:国防工业出版社,2007.
[115]飞思科技产品研发中心. MATLAB6.5辅助小波分析与应用[M].北京:电子工业出版社,2003.
[116] Cybenko G. Approximation by superpositions of a sigmoid function[J]. Math Control,Signal and Systems,1989,2(4):201-205.
[117] ZHANGQing2hua, Benveniste Albert. Wavelet Networks[J]. IEEE Trans on NN,1992,3(6):311-31
[118]秦前清,杨宗凯.实用小波分析[M].西安:西安电子科技大学出版社,1995.
[119]崔锦泰.程正兴译.小波分析导论[M].西安:西安交通大学出版社,1995.
[120] Bakshi B R, Stephanopoulos G.Wavenet: a multiresolution hierachical neural network withlocalized learning [J]. AICHEJ,1993,39(1):57-81.
[121] Daubechies I.Orthonormal base of compactly supported wavelets [J]. Communications onPure and Applied Mathematics,1988,41(2):909-996.
[122]武君.河流水质模拟预测的常用方法研究与新方法探索——以淮河安徽段为例[D].合肥:合肥工业大学,2005.
[123]钟锦.基于演化博弈的淮河流域水环境管理研究[D].合肥:合肥工业大学,2008.
[124]党建武.神经网络技术及应用[M].北京:中国铁道出版社,2000.
[125]丛爽.面向MATLAB工具箱的神经网络理论与应用[M].合肥:中国科学技术大学出版社,2003.
[126] Tony H. Quantum computing: an introduction[J]. Computing&Control Engineering Journal,1996;10(3):105-112.
[127]雷英杰,张善文,李续武,等. Matlab遗传算法工具箱及应用[M].西安:西安电子科技大学出版社,2005.
[128]周明,孙树栋.遗传算法原理及应用[M].国防工业出版社,2002.
[129]周正武,丁同梅,田毅红,等. Matlab遗传算法优化工具箱(GAOT)的研究与应用[J].机械研究与应用,2006,6(19):69-71.
[130] Tony H. Quantum computing: an introduction[J]. Computing&Control Engineering Journal,1996;10(3):105-112.
[131] Han K H, Park K H, Lee C H,et al. Parallel quantum-inspired genetic algorithm forcombinatorial optimization problems[A]. Proceedings of IEEE International Conference onEvolutionary Computation[C]. Piscataway: IEEE Press,2001:1442-1429.
[132]李士勇,李盼池.量子计算与量子优化算法[M].哈尔滨:哈尔滨工业大学出版社,2009.