基于鸡群算法优化相关向量机的混凝土坝变形预报模型

针对混凝土坝变形监控模型中大坝变形与环境影响因素之间的复杂非线性问题,为提升大坝变形监控模型的预报能力,提出了一种基于鸡群算法(CSO)优化相关向量机(RVM)的混凝土坝变形预报模型。考虑到相关向量机核函数参数的选取直接影响其回归分析性能,采用鸡群算法对其核函数参数进行寻优处理。据此,构造了基于鸡群算法优化的相关向量机模型,进而提升相关向量机的预报精度和泛化能力。以某混凝土坝长期变形监测资料分析表明,基于鸡群算法优化的相关向量机模型预报可有效挖掘大坝变形与环境因素间复杂的非线性函数关系,相比传统的相关向量机模型,该模型的拟合与预报精度更优,有效验证了所提方法的合理性与有效性,为大坝变形分析与预测提供新的模型方法。

Chicken swarm optimization algorithm-based optimization ofrelevance vector machine model for concrete dam deformation prediction

Aiming at the complicated nonlinear problems between the dam deformation and the environmental influencing factors in the concrete dam monitoring model，a chicken swarm optimization( CSO) algorithm-based optimization of relevance vector machine( RVM) model for concrete dam deformation monitoring is put forward herein，so as to enhance the prediction capacity of the dam deformation monitoring model. Under the consideration of that the selection of the kernel function parameter of the relevance vector machine directly affects the performance of its regressive analysis，its kernel function parameter is optimized through chicken swarm optimization( CSO) algorithm. Hereby，the relevance vector machine( RVM) model that is optimized on the basis of the chicken swarm optimization( CSO) algorithm is constructed，and then the prediction accuracy and generalization capacity of the relevance vector machine( RVM) are enhanced. The analysis made on the long-term deformation monitoring data of a concrete dam shows that the prediction from therelevance vector machine ( RVM) model optimized through the chicken swarm optimization ( CSO) algorithm can effectively find out the nonlinear functional relationship between dam deformation and the environment factors，while the fitting and prediction accuracies of the model are more optimal，thus not only the reasonability and effectiveness of the method proposed herein are effectively certified，but a new modeling method is also provided for the dam deformation analysis and prediction.