基于AdaBoost集成的WPSO-RBF大坝变形监控模型
|
摘要
变形监测是大坝安全监测的必设项目,由于影响因子众多,常利用神经网络(如BP,RBF等)进行参数选取和模型建立。传统的径向基函数(RBF)神经网络因网络结构简单、收敛速度快而被广泛运用,但其在预测中易陷入局部最优且参数选取不当会对其收敛性产生影响。因此,首先利用动态权重粒子群算法(WPSO)对RBF神经网络的3个参数(隐含层基函数的中心c、宽度d及隐含层到输出层的权值w)进行优化,建立基于WPSO-RBF的大坝变形监控模型,然后将WPSO-RBF模型作为弱分类器,采用Ada Boost算法进行集成,建立基于WPSO-RBF-Ada Boost的大坝变形监控模型。将该模型运用到工程实例中,实例结果显示该模型具有收敛速度快、分类精度高、泛化能力好,可建立较优的大坝变形监控模型。
Deformation monitoring is a requisite for dam safety monitoring. Due to a large number of factors,neural networks such as back propagation( BP) and radial basis function( RBF) are often used for parameters selection and model establishment,of which RBF has been widely employed on account of its simple network structure and rapid convergence. Nonetheless,local optimality and inappropriate selection of parameters will exert great impact on the convergence rate. In view of this,the Particle Swarm Optimization with Inertia Weight( referred to as WPSO)is adopted to optimize three parameters of RBF( central value c of hidden layer base function parameter,width d and connection weight w between hidden layer and output layer parameter). In subsequence,the WPSO-RBF model is integrated as a weaker classifier by Ada Boost algorithm,hence establishing a WPSO-RBF-Ada Boost model for dam deformation monitoring. The model is applied to practical engineering,and results suggest that the present model is of fast convergence,high classification precision and good generalization ability.
Deformation monitoring is a requisite for dam safety monitoring. Due to a large number of factors,neural networks such as back propagation( BP) and radial basis function( RBF) are often used for parameters selection and model establishment,of which RBF has been widely employed on account of its simple network structure and rapid convergence. Nonetheless,local optimality and inappropriate selection of parameters will exert great impact on the convergence rate. In view of this,the Particle Swarm Optimization with Inertia Weight( referred to as WPSO)is adopted to optimize three parameters of RBF( central value c of hidden layer base function parameter,width d and connection weight w between hidden layer and output layer parameter). In subsequence,the WPSO-RBF model is integrated as a weaker classifier by Ada Boost algorithm,hence establishing a WPSO-RBF-Ada Boost model for dam deformation monitoring. The model is applied to practical engineering,and results suggest that the present model is of fast convergence,high classification precision and good generalization ability.
引文
[1]吴中如.水工建筑物安全监控理论及其应用[M].北京:高等教育出版社,2003.
[2]SU Huai-zhi,HU Jiang,WU Zhong-ru.A Study of Safety and Early-warning Method for Dam Global Behavior[J].Structural Health Monitoring,2012,11(3):269-279.
[3]王凌.智能优化算法及其应用[M].北京:清华大学出版社,2001.
[4]PARK J,SANDBERG I W.Universal Approximation Using Radial-Basis-Function Networks[J].Neural Computation,1991,3(2):246-257.
[5]张辉,柴毅.一种改进的RBF神经网络参数优化方法[J].计算机工程与应用,2012,48(20):146-149,157.
[6]胡德华,郑东健,付浩雁.Ada Boost-BP模型在大坝变形预测中的应用[J].三峡大学学报(自然科学版),2015,37(5):5-8.
[7]SHI Y H,EBERHART R C.Empirical Study of Particle Swarm Optimization[C]∥Proceedings of the 1999 Congress on Evolutionary Computation-CEC 99.Washington,DC,USA,July 6-9,1999:1945-1950.
[8]黄剑竹.改进RBF-Ada Boost模型及其在年径流预测中的应用[J].人民珠江,2015,(1):32-36.
[9]胡昊,魏继红.基于PSO-RBF的边坡安全系数预测[J].河南科学,2016,34(5):747-751.
[10]CHEN S,HONG X,LUK B L,et al.Non-linear System Identification Using Particle Swarm Optimization Tuned Radial Basis Function Models[J].International Journal of Bio-inspired Computation,2009,1(4):246-258.
[11]周驰,高海兵,高亮,等.粒子群优化算法[J].计算机应用研究,2003,20(12):7-11.
[12]SHI Y H,EBERHART R C.A Modified Particle Swarm Optimizer[C]∥Proceedings of IEEE International Conference on Evolutionary Computation.Anchorage Alaska,USA,May 4-9,1998:69-73.
[13]刘建华.粒子群算法的基本理论及其改进研究[D].长沙:中南大学,2009:22-46.
[14]卢金娜.基于优化算法的径向基神经网络模型的改进及应用[D].太原:中北大学,2015.
[15]王小川,史峰,郁磊.MATLAB神经网络43个案例分析[M].北京:北京航空航天大学出版社,2016:42.
[16]蒋明,郑东健,张凯,等.FA-SVM模型在大坝变形预测中的应用[J].水力发电,2015,41(4):79-81,99.
[17]乔俊飞,韩红桂.前馈神经网络分析与设计[M].北京:科学出版社,2013.
[2]SU Huai-zhi,HU Jiang,WU Zhong-ru.A Study of Safety and Early-warning Method for Dam Global Behavior[J].Structural Health Monitoring,2012,11(3):269-279.
[3]王凌.智能优化算法及其应用[M].北京:清华大学出版社,2001.
[4]PARK J,SANDBERG I W.Universal Approximation Using Radial-Basis-Function Networks[J].Neural Computation,1991,3(2):246-257.
[5]张辉,柴毅.一种改进的RBF神经网络参数优化方法[J].计算机工程与应用,2012,48(20):146-149,157.
[6]胡德华,郑东健,付浩雁.Ada Boost-BP模型在大坝变形预测中的应用[J].三峡大学学报(自然科学版),2015,37(5):5-8.
[7]SHI Y H,EBERHART R C.Empirical Study of Particle Swarm Optimization[C]∥Proceedings of the 1999 Congress on Evolutionary Computation-CEC 99.Washington,DC,USA,July 6-9,1999:1945-1950.
[8]黄剑竹.改进RBF-Ada Boost模型及其在年径流预测中的应用[J].人民珠江,2015,(1):32-36.
[9]胡昊,魏继红.基于PSO-RBF的边坡安全系数预测[J].河南科学,2016,34(5):747-751.
[10]CHEN S,HONG X,LUK B L,et al.Non-linear System Identification Using Particle Swarm Optimization Tuned Radial Basis Function Models[J].International Journal of Bio-inspired Computation,2009,1(4):246-258.
[11]周驰,高海兵,高亮,等.粒子群优化算法[J].计算机应用研究,2003,20(12):7-11.
[12]SHI Y H,EBERHART R C.A Modified Particle Swarm Optimizer[C]∥Proceedings of IEEE International Conference on Evolutionary Computation.Anchorage Alaska,USA,May 4-9,1998:69-73.
[13]刘建华.粒子群算法的基本理论及其改进研究[D].长沙:中南大学,2009:22-46.
[14]卢金娜.基于优化算法的径向基神经网络模型的改进及应用[D].太原:中北大学,2015.
[15]王小川,史峰,郁磊.MATLAB神经网络43个案例分析[M].北京:北京航空航天大学出版社,2016:42.
[16]蒋明,郑东健,张凯,等.FA-SVM模型在大坝变形预测中的应用[J].水力发电,2015,41(4):79-81,99.
[17]乔俊飞,韩红桂.前馈神经网络分析与设计[M].北京:科学出版社,2013.