| 基于遗传算法因素筛选的BP神经网络在软土路基沉降数据处理中的应用 |
| |
| 引用本文： | 王江荣,赵睿,袁维红,任泰明.基于遗传算法因素筛选的BP神经网络在软土路基沉降数据处理中的应用[J].矿山测量,2016(5):87-90. |
| |
| 作者姓名： | 王江荣 赵睿 袁维红 任泰明 |
| |
| 作者单位： | 1. 兰州石化职业技术学院 信息处理与控制工程系 兰州 730060;2. 兰州石化职业技术学院 土木工程系,甘肃 兰州,730060 |
| |
| 基金项目： | 兰州市科学技术局计划项目(兰财建发[2015]85号)，兰州石化职业技术学院科技资助项目(院发[2015]69号)，甘肃省科技厅计划项目(1204GKCA004)，甘肃省财政厅专项资金立项资助(甘财教[2013]116号) |
| |
| 摘 要： | 软土路基沉降与其影响因素之间存在着非线性关系,因输入自变量较多,用神经网络建模容易出现过拟合现象,导致网络模型预测精度降低。针对这个问题,提出用遗传算法对输入自变量进行压缩降维处理,同时对网络模型的权值和阈值进行优化。实例仿真表明:经降维和权值及阈值优化的BP网络具有较高的精度;预测效果优于GRNN网络模型和单纯BP网络模型;用于软土路基沉降预测是可行的。
|
| 关 键 词： | 软土路基沉降 BP神经网络 遗传算法 因素筛选 沉降预测 |
Application of BP neural network Factor screening based onselected by genetic algorithmin data processing of soft soil subgradesub-grade settlement |
| |
| Abstract: | There is a nonlinear relationship between the settlement of soft soil subgradesub-grade and its influencing factors,. bBecause of the numerous input variables are more, the neural network modeling is prone to over fitting phenomena , which leads to the decrease of the prediction accuracy of the network model. To solve this problem, a genetic algorithm is proposed to reduce the dimension of the input variables, and the weights and thresholds of the network model are optimized. The simulation results show that the BP network has a high precisionaccuracy, and the predictive effect is better than that of the GRNN network model and the BP network model,. andTherefore it is feasi-ble to predict the settlement of soft soil subgrade. |
| |
| Keywords: | soft soil subgrade settlement BP neural network genetic algorithm factor selection settlement predic-tion |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
