论文部分内容阅读
水利工程实施的效果可以用增泄水量的多少来评价,文中构建了一种计算增泄水量的计算模型。以径向基函数(RBF)作为核函数,建立了以上游来水量、中游GDP增长、人口增长、降水量为输入,下游下泄水量为输出的支持向量机计算模型,为了提高支持向量机的预测精度,利用混沌粒子群算法(CPSO)的全局寻优特性进行支持向量机(SVM)的参数辨识,克服了人工选取的不足。以黑河流域(1990~2007年)18年的数据样本集作为训练样本,将后5年(2008~2012年)的数据样本集作为检验样本,选择参数如下:C=100,ε=0.001,σ=14。通过支持向量机模型计算的最大相对误差为8.01%,平均相对误差为6.50%。结果表明:文中建立的基于CPSO-SVM的增泄水量计算模型具有很好的效果,可以用于对水利工程实施效果的评价。
The effect of water conservancy project implementation can be used to increase the amount of discharge water to evaluate, in this paper, a calculated calculation of water discharge calculation model. Based on the Radial Basis Function (RBF) as the kernel function, a support vector machine (SVM) model was established for the input of upstream water flow, middle reaches of GDP growth, population growth, precipitation as the input and the downstream discharge as output. In order to improve the prediction of SVM Accuracy, and the global optimization of chaos particle swarm optimization (CPSO) is used to identify the parameters of support vector machine (SVM), which overcomes the shortcomings of manual selection. Data samples of 18 years in Heihe River Basin (1990-2007) were used as training samples, and data samples of the last 5 years (2008-2012) were used as test samples. Parameters were selected as follows: C = 100, ε = 0.001, σ = 14. The maximum relative error calculated by support vector machine model is 8.01%, and the average relative error is 6.50%. The results show that the CPSO-SVM-based calculation model for water discharge increases the effectiveness of the model and can be used to evaluate the effect of water conservancy projects.