摘要
针对紧邻大型深基坑的地铁隧道因其变形影响因素复杂、变形控制严格而难以准确预测其沉降变形的问题,文章引入对小样本、复杂、非线性数据具有优越预测性能的支持向量机理论,并利用蚁群优化算法搜索支持向量机最优参数组合,建立了优化的支持向量机预测模型。应用该模型对南京市地铁1号线某段隧道的预测结果表明,该模型预测精度高,能够准确反映隧道变形趋势,可以满足紧邻大型深基坑地铁隧道沉降预测的要求。
Due to the complex factors of and strict requirements for deformation control, it is difficult to accurate-ly predict the settlement of a metro tunnel adjacent to a deep large-scale foundation pit. In this paper, the Sup-port Vector Machine （SVM） theory is introduced for its excellent prediction capabilities regarding small sample, complex, and nonlinear data, and Ant Colony Optimization （ACO） was adopted to obtain an optimized SVM pre-diction model by selecting the most appropriate parameter combination. Based on the settlement prediction of a tunnel on Nanjing Metro Line 1, this model meets the requirements of settlement prediction for metro tunnels ad-jacent to deep large-scale foundation pits because of its sound prediction accuracy and accurate reflection of de-formation trends.
出处
《现代隧道技术》
EI
北大核心
2014年第6期94-100,共7页
Modern Tunnelling Technology
关键词
地铁隧道
沉降预测
深基坑
支持向量机
蚁群优化算法
参数优化
Metro tunnel Settlement prediction Deep foundation pit Support Vector Machine （SVM） Ant Colony Optimization （ACO） Parameter optimization
