首页 | 本学科首页   官方微博 | 高级检索  
     检索      

地震静校正全局最优化问题的求解
引用本文:井西利,杨长春,李幼铭,白彦彬.地震静校正全局最优化问题的求解[J].地球物理学报,2002,45(5):707-713.
作者姓名:井西利  杨长春  李幼铭  白彦彬
作者单位:中国科学院地质与地球物理研究所,北京,100101
摘    要:针对地震静校正存在的非线性和多参数性等问题,本文提出了一种全局优化的剩余静校正方法.在模型更新规则上,把模拟退火方法和均匀优化设计方法结合在一起,使模型的更新更为合理,从而加快了搜寻全局最优解的速度.同时提出了模型的不同分量的温度参数和退火过程的选取方式,把温度参数的选取与地震剖面的能量联系起来,使温度参数的选取具有自适应的特点.该方法克服了常规模拟退火方法所具有的寻优空间不均匀以及退火参数需通过多次试验选取的缺陷.通过实际地震资料的计算证明,本文提出的方法对地震静校正问题合理而有效.

关 键 词:静校正  模拟退火  均匀设计  非线性反演  
文章编号:0001-5733(2002)05-0707-07
收稿时间:2001-05-22

A GLOBAL OPTIMIZED ALGORITHM FOR SEISMIC RESIDUAL STATICS CORRECTIONS
JING XI LI\ \ YANG GHENG CHUN,LI YOU MING\ \ BAI YAN BIN.A GLOBAL OPTIMIZED ALGORITHM FOR SEISMIC RESIDUAL STATICS CORRECTIONS[J].Chinese Journal of Geophysics,2002,45(5):707-713.
Authors:JING XI LI\ \ YANG GHENG CHUN  LI YOU MING\ \ BAI YAN BIN
Institution:Institute of Geology and Geophysics, Chinese Academy of Sciences, Beijing 100101, China
Abstract:A global optimized algorithm is developed in this study for dealing with the problems of nonlinearity and multiparameter characteristics of the residual statics correction method. Integration of the simulated annealing approach with the uniform design optimized method makes the updating of the model more reasonable and therefore,increases the speed of searching for the global optimimzed solution. A procedure is proposed for selecting the temperature parameters of different model components and the process of decreasing the temperature. Here the selection of the temperature parameter is correlated with seismic energy, which makes the algorithm having a self-adapted characteristic. This new algorithm overcomes several shortcomings of the conventional simulated annealing method. Its application to seismic data shows that the algorithm developed in this study is more efficient than the conventional simulated annealing method.
Keywords:Statics corrections  Simulated annealing  Uniform design  Nonlinear inversion  
本文献已被 CNKI 维普 万方数据 等数据库收录!
点击此处可从《地球物理学报》浏览原始摘要信息
点击此处可从《地球物理学报》下载免费的PDF全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号