摘要
由于神经网络具有较好的自组织和自适应特点,已被广泛应用于基坑变形预测领域。但神经网络中节点的权值最优化难以处理。为进一步提高BP神经网络性能,实现准确、快速预测基坑变形的目的,可将蚁群算法作为BP神经网络的学习算法,建立一种新的蚁群神经网络预测模型。实例表明,基于蚁群—BP神经网络的基坑变形预测方法与传统的BP神经网络预测方法相比,具有较强的自适应能力,取得了较好的效果。
The neural network is widely used in the foundation deformation prediction field for its good characteristics of adaptive self-organizing. Node in the neural network weights optimization is difficult to handle, in order to further improve the performance of the.BP neural network, to achieve accurate and quick prediction of the foundation deformation, the ant colony algorithm is used as the learning algorithm of BP neural network and a new ant colony neural network prediction model is established. The example shows that the foundation deformation based on ant colony-BP neural network prediction methods compared with traditional BP neural network prediction method, with a strong adaptive capacity and better results.
出处
《现代测绘》
2012年第6期13-14,27,共3页
Modern Surveying and Mapping
基金
南京市科技计划项目(201101069)
江苏省测绘科研项目(JSCHKY201108)
江苏省建设厅科技项目(JS2011JH23)
关键词
基坑变形预测
BP神经网络
蚁群算法
deformation prediction of pit foundation
BP neural network
ant colony algorithm
