基于改进ABC-BP的大坝变形监控模型研究

Research on dam deformation monitoring model based on improved ABC- BP

针对传统大坝变形监控模型的不足,在对人工蜂群(ABC)算法给予改进的基础上,开展了基于人工蜂群(ABC)与BP神经网络的大坝变形监控模型建模原理、实现方法以及工程算例分析研究。通过引进自适应比例和平均欧式距离,克服了标准人工蜂群算法易陷入局部最优的缺点;进而利用改进后的人工蜂群算法,对BP神经网络的初始权值与阈值进行寻优。算例分析表明,将改进后的人工蜂群算法与BP神经网络技术相结合,并用于大坝变形监控模型的构建,有效提升了模型的拟合和预报能力。

Aimed at the shortcomings of traditional dam monitoring model,the research on combinative application of Artificial Bee Colony algorithm and BP neural network in dam deformation monitoring were carried out,in terms of modeling principle,realization methodology and engineering case,on the basis of improving the traditional Artificial Bee Colony（ ABC） algorithm. The adaptive proportion and average Euclidean distance were introduced into standard Artificial Bee Colony algorithm for overcoming the local optimal. Then,the initial weights and thresholds of BP neural network were optimized by the improved Artificial Bee Colony algorithm. Engineering application results showed that the combination of the improved Artificial Bee Colony algorithm and BP neural network technology can be used to establish dam deformation monitoring model,and effectively improve the model＇s fitting and forecasting ability.