| 摘要: |
| 【目的】 针对常规大坝变形监测回归模型中存在的因子多重相关性干扰和模型拟合效果欠佳问题，进行偏回归模型优化方法研究。【方法】 将改进的遗传算法引入大坝变形监测偏回归模型，利用遗传算法强大的自适应全局优化搜索功能，对偏最小二乘回归模型进行优化，建立了基于改进遗传算法-偏最小二乘回归的大坝变形监测模型。【结果】 工程实例研究与对比分析表明，改进遗传算法-偏最小二乘回归模型在一定程度上改善了原偏回归模型存在的拟合效果不佳的问题。【结论】 改进遗传算法-偏最小二乘回归模型具有较好的拟合与预测能力，有较强的工程实用性。 |
| 关键词: 改进遗传算法 偏最小二乘法 回归模型 大坝变形监测 |
| DOI： |
| 分类号: |
| 基金项目:国家自然科学基金项目（50779051）；水资源与水电工程科学国家重点实验室开放基金项目（2007B037）；陕西省教育厅专项科研计划项（07JK354）；西安理工大学科学研究基金项目（106 210509） |
|
| Monitoring model for dam deformation based on partial least-squared regression and improved genetic algorithm |
|
|
| Abstract: |
| 【Objective】 Aiming at the problem of the under-fitting and multicollinearity in general dam safety monitoring model,an optimization of partial regression was put forward.【Method】 The improved genetic algorithm was introduced in the modeling of partial least-squared regression in dam deformation monitoring to optimize the partial least-squares regression model by using its powerful adaptive global optimization search function,and the monitoring model established for dam deformation based on partial least-squared regression and improved genetic algorithm.【Result】 Project instance studies and correlation analysis show that the model based on partial least-squared regression and improved genetic algorithm improved the under-fitting phenomenon of the original model of partial regression to some extent.【Conclusion】 The model based on partial least-squared regression and improved genetic algorithm has good simulating effect and forecasting precision and strong engineering practicability. |
| Key words: improved genetic algorithm partial least-squared regression model dam deformation monitoring |
