共查询到20条相似文献,搜索用时 31 毫秒
1.
基于人工鱼群算法和模糊C-均值聚类的洪水分类方法 总被引:2,自引:0,他引:2
为了克服模糊C-均值聚类(FCM)算法依赖初值的缺点,引入人工鱼群算法(AFS)建立一种新的聚类算法,应用于洪水分类研究。该算法将聚类中心看作食物源,通过样本抽样产生初始鱼群,利用人工鱼群算法能全局寻优和快速收敛的特点,得到一个较优的初始聚类结果,再使用FCM算法进行局部搜索,以避免因初值选取不当,而有可能陷入局部最小的缺陷。该方法应用于对西江流域洪水资料的分析结果表明,新算法具有比FCM算法更好的性能表现,使得到的分类结果更加准确合理。 相似文献
2.
投影寻踪和人工鱼群算法的洪水分类 总被引:3,自引:1,他引:2
由于洪水具有的复杂性,从而增加了洪水类别划分的难度。通过人工鱼群算法优化投影寻踪模型,提出了一种新的聚类算法,并将其应用于洪水分类研究。由于鱼群觅食与聚类有着天然的相似性,将聚类中心看作食物源,通过样本抽样产生初始鱼群,利用人工鱼群算法进行全局寻优。为了说明该算法用于解决洪水分类问题的可行性和优越性,采用长江流域的重要控制站—宜昌站为研究对象,根据历史文献提供的相关数据进行分析,与实际洪水情况及前人的分类结果相对比。从分类结果来看,能说明采用投影寻踪与人工鱼群算法的可行性,通过性能比较,则进一步显示出该算法的优越性。 相似文献
3.
利用混合型模糊聚类分析方法求出模糊聚类初始相对隶属度矩阵和模糊聚类中心,根据熵值算法求得各指标的客观权重值,再根据最大隶属度原则和权重因子可清晰地给出样本的分类。实例应用结果表明:该方法具有良好的应用性,可以克服权重赋值的人为干扰。 相似文献
4.
5.
洪水聚类有效性分析 总被引:1,自引:0,他引:1
在聚类分析中,聚类数是一个非常重要的参数,最佳聚类数的确定问题是聚类分析研究的热点之一。在模糊聚类迭代模型的基础上,首先提出了基于类间相关系数的聚类有效性指标确定最佳聚类数,并给出了最佳聚类数的确定步骤。随后以IRIS和Ruspini数据集作为分析样本验证了所提聚类有效性指标的有效性。以双牌水库的45场典型洪水过程和碧流河水库的13场典型洪水过程为例,重点进行了洪水聚类分析,进一步验证了所提公式的有效性。碧流河水库洪水聚类,遵循了成因分析和聚类分析相结合的原则,其目的是确定各类主要天气系统的典型代表过程,以便进行水库防洪分类预报调度设计。 相似文献
6.
7.
为解决流域历史洪水资料有限引起的洪水预报模型模拟精度不高的问题。以A水库为研究对象,利用K均值聚类方法对典型洪水进行聚类,分析降雨雨强、降雨中心和天气系统等水文影响因子,通过遗传优化算法计算汇流模型的各类参数,利用粗糙集方法挖掘影响因子与时段汇流模式间的关系,建立了基于信息融合与识别的洪水时段分类预报。结果表明:(1)选取的4场典型洪水通过分类预报方法计算得到的洪峰流量绝对误差与相对误差的绝对值分别为第一场9.01 m3/s、2.95%,第二场116.46 m3/s、6.78%,第三场30.92 m3/s、17.55%,第四场6.12 m3/s、1.86%;(2)洪水分类预报模型的模拟精度较传统预报方法更高,不同典型洪水的确定系数均在0.8以上。研究结果可为洪水资料较少的华北等地区的洪水时段分类预报提供参考和借鉴。 相似文献
8.
9.
根据河南省1960-2017年19个气象站的逐日降雨数据计算标准化降水指数(SPI),结合模糊C均值聚类法、小波分析,从月、年际和周期维度上研究近57年来河南省干旱的时空变化特征。结果表明:所有气象站点可以分成3个区域,分别为豫东南、豫北、豫西地区;全区在1966、1986、1997年发生持续干旱,豫东南持续干旱时间最长为10个月且在夏季易旱,豫北持续干旱时间最长为9个月且易发生特旱,豫西持续干旱时间段最多但尺度最长为7个月;干旱频率主要分布在豫北,豫西次之,豫东南最小,但干旱频率都在30%以上,且20世纪90年代后干旱频率在各区之间分布比较均匀;豫东南、豫北、豫西分别呈现23~25、20~22、15~17 a易干旱的主周期。 相似文献
10.
本文以水资源平衡为研究基础,引进了灌区水资源联合调控的人工鱼群算法,建立了地下水与地表水间的联合调控模型。得出了不同水平年50%保证率和75%保证率下的灌溉水资源分配结果,为今后灌区水资源联合调控提供参考。 相似文献
11.
12.
In recent years, evolutionary techniques have been widely used to search for the global optimum of combinatorial non-linear non-convex problems. In this paper, we present a new algorithm, named fuzzy Multi-Objective Particle Swarm Optimization (f-MOPSO) to improve conjunctive surface water and groundwater management. The f-MOPSO algorithm is simple in concept, easy to implement, and computationally efficient. It is based on the role of weighting method to define partial performance of each point (solution) in the objective space. The proposed algorithm employs a fuzzy inference system to consider all the partial performances for each point when optimizing the objective function values. The f-MOPSO algorithm was compared with two other well-known MOPSOs through a case study of conjunctive use of surface and groundwater in Najafabad Plain in Iran considering two management models, including a typical 12-month operation period and a 10-year planning horizon. Overall, the f-MOPSO outperformed the other MOPSO algorithms with reference to performance criteria and Pareto-front analysis while nearly fully satisfying water demands with least monthly and cumulative groundwater level (GWL) variation. The proposed algorithm is capable of finding the unique optimal solution on the Pareto-front to facilitate decisions to address large-scale optimization problems. 相似文献
13.
Reconstruction and/or modification of an already existing fuzzy model with new data may improve system performances. As new
data become available, adjusting the existing fuzzy rule-based model may present a challenging alternative to full model reconstruction.
In this paper a fuzzy rule-based control model using a Takagi–Sugeno fuzzy system is presented and a model modification algorithm
is developed which improves the performance of the initial model as new data become available. Proposed approach is applied
to a flood flow forecasting case example and the results are compared with those forecasted using initially available and
reconstructed models. Results show that the modified model outperforms the initial FRB model. Reconstructed model performs
slightly better than the modified model; however, the reconstruction may not be justified in a real time flood forecasting
system, considering the limitations on the available lead time. 相似文献
14.
Mohammad Karamouz Ozeair Abesi Ali Moridi Azadeh Ahmadi 《Water Resources Management》2009,23(9):1743-1761
In this paper, two optimization models are presented. The first model is developed to determine economical combination of
permanent and emergency flood control options and the second one is used to determine the optimal crop pattern along a river
based on the assigned flood control options by the first optimization model. The optimal combination of flood protection options
is determined to minimize flood damages and construction cost of flood control options along the river using the genetic algorithm
(GA) optimization model. In order to consider the effects of flood control options on hydraulic characteristics of flow, two
hydrological routing models for the reservoir and the river are coupled with the optimization model. Discharge–elevation and
elevation–damage curves obtained based on separate hydraulic simulations of the river are used for flood damage calculations
in the optimization model. The parameters of a hydrologic river routing model are also calibrated using the developed hydraulic
model results. The proposed model is applied to the Kajoo river in the south-eastern part of Iran. The results demonstrate
an economical integration of permanent and emergency flood control options along the river which include minimum expected
value of damages related to floods with different return periods and construction cost of flood control options. Finally the
resulting protection scheme is used for land use planning through identifying the optimal crop mix along the river. In this
approach, the objective function of the optimization model is an economic function with a probabilistic framework to maximize
the net benefit of agricultural activities. The study exhibits the importance of floodplain management and land use planning
to achieve the development goals in the river basins. 相似文献
15.
16.
人工神经网络及其在水科学研究中的应用 总被引:4,自引:0,他引:4
概述了人工神经网络(ANN)理论的基本原理及其有关概念,介绍了人工神经网络中的BP网络模型和学习算法,回顾总结了近年来ANN技术在水科学研究诸领域应用的主要成果,并指出今后应用研究的方向。 相似文献
17.
18.
基于自由搜索的灌区优化配水模型研究 总被引:6,自引:0,他引:6
在求解优化配水问题时,遗传算法和粒子群算法是经常被采用的优化算法,然而在优化过程中可能出现早熟,从而会影响到配水模型的有效性。为此,本文引入一种新的基于动物群体的优化算法——自由搜索(FreeSearch)算法,并以实现单位灌溉水量的净收益最大为目标构建了灌区优化配水模型,设自由搜索中动物探查行走时的位置分量为模型寻优参数,应用自由搜索算法对模型进行优化求解。实例应用结果表明:与加速遗传算法和标准粒子群算法求解优化配水模型的结果相比,本文建立优化配水模型能为灌区提供更合理的优化配水方案,可使整个灌区及单位灌溉水量的净收益都获得显著增长。 相似文献
19.
提出了基于单纯形微粒群优化算法(SMPSO)直接求解作物水分生产函数Jensen模型参数的方法。通过计算分析,用传统方法计算而得Jensen模型的相对产量与实际相对产量的相关指数为0.825,而用SMPSO(PSO)计算而得的相关指数为0.932;SMPSO寻优仅需要0.3973s,而PSO需要1.0783s。 相似文献
20.