| 基于基因表达式编程算法的波浪透射预测 |
| |
| 引用本文： | 刘涛,冯曦,冯卫兵,张宸豪,陆杨.基于基因表达式编程算法的波浪透射预测[J].海洋工程,2021,39(1):133-141. |
| |
| 作者姓名： | 刘涛 冯曦 冯卫兵 张宸豪 陆杨 |
| |
| 作者单位： | 海岸灾害及防护教育部重点实验室(河海大学), 江苏 南京 210098;河海大学 港口海岸与近海工程学院, 江苏 南京 210098;南京水利科学研究院, 江苏 南京 210029 |
| |
| 基金项目： | 国家重点研发计划课题（2017YFC0405206）；国家自然科学基金（51709091）；江苏省自然科学基金（BK20170874）；中央高校基金（2017B005） |
| |
| 摘 要： | 准确预测波浪透射对于维护港内水域平稳、保障港内船舶稳定具有重要意义。基于567组透浪试验数据,采用基因表达式编程(gene expression programming,简称GEP)算法预测波浪透射。主要研究内容包括:确定GEP算法的最优输入变量组合;建立透浪系数与最优组合变量的定量关系;探究GEP算法的预测精度随训练组数变化的规律;并对输入变量进行了敏感性分析。研究结果表明,GEP算法的最优输入变量组合为深水波陡、相对堤宽和相对水深;训练组数较少时,GEP算法的预测精度不高,当训练组数提高至300组,预测的精度已经达到较高水平,且精度随着训练组数的继续增加提高不大; GEP算法的预测精度远远高于前人的经验公式;相较于相对堤宽和相对水深,深水波陡对波浪透射影响更为显著。本研究表明,GEP算法可作为一种新的方法研究波浪透射,为后续研究与应用提供参照。
|
| 关 键 词： | 透浪系数 基因表达式编程 敏感性分析 深水波陡 |
| 收稿时间： | 2020/3/9 0:00:00 |
Prediction of wave transmission using gene expression programming |
| |
| LIU Tao,FENG Xi,FENG Weibing,ZHANG Chenhao,LU Yang.Prediction of wave transmission using gene expression programming[J].Ocean Engineering,2021,39(1):133-141. |
| |
| Authors: | LIU Tao FENG Xi FENG Weibing ZHANG Chenhao LU Yang |
| |
| Institution: | Key Laboratory of Coastal Disaster and Protection(Hohai University), Ministry of Education, Nanjing 210098, China;College of Harbor, Coastal and Offshore Engineering, Hohai University, Nanjing 210098, China; Nanjing Hydraulic Research Institute, Nanjing 210029, China |
| |
| Abstract: | Accurate prediction of wave transmission is of great significance for maintaining the stability of waters and ensuring the stability of ships in the harbor. Based on 567 sets of experimental data, this study used gene expression programming (GEP) algorithm to predict wave transmission. The optimal combination of the input variables for the GEP method was firstly determined. Then a quantitative relationship was established between wave transmission coefficient and the optimal combination of the input variables. In the last, multiple sensitivity tests were conducted on the accuracy of prediction for the GEP method with the number of training groups and other parameters. Results indicate that deep-water wave-steepness, relative width of the breakwater and relative water depth together are the optimal combination. The predictive ability of the GEP method was low when the training number was small but increased fast as long as the training number was below 300. When the training number exceeded 300, the improvement of the accuracy score slows down. The GEP method predicting wave transmission outperformed a number of existing formulas. The deep-water wave-steepness had a more significant effect on wave transmission in comparison with the relative width of the breakwater and relative water depth. The research shows that the GEP algorithm can be used as a new method to investigate wave transmission, which can provide a reference for subsequent research and application of this new methodology. |
| |
| Keywords: | wave transmission coefficient gene expression programming sensitivity analysis deep-water wave-steepness |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《海洋工程》浏览原始摘要信息 |
| 点击此处可从《海洋工程》下载免费的PDF全文 |
|
