基于BP-GA算法的微震源反演研究
详细信息 本馆镜像全文    |  推荐本文 | | 获取馆网全文
摘要
针对传统非线性反演方法易陷入局部最优,收敛速度慢等问题,研究了一种基于BP神经网络和遗传算法的微震源定位反演方法。重点论述了BP-GA混合算法的理论基础与模型建立,并结合实例进行了3层介质模型微震源定位反演仿真。
For the defects of traditional nonlinear inversion methods,such as easy to fall into local optimum resolution and slow rate of convergence,a micro-seismic source positioning inversion method based on BP-GA algorithm has been proposed.The design model and theory of the BP-GA hybrid algorithm is discussed in this paper.It also describes the method of data analysis and processing in Matlab with instances.
引文
[1]Wei Chao,Li Xiaogan,Zheng Xiaofong.The group search-based parallel algorithm for the serial monte carlo inversionmethod[J].Applied Geophysics,2010,7(2):127-134.
    [2]田玥,陈晓非.地震定位研究综述[J].地球物理学进展,2002,17(1):147-155.
    [3]巩思园,窦林名,马小平,等.煤矿矿震定位中异向波速模型的构建与求解[J].地球物理学报,2012,55(5):1757-1763.
    [4]陈霄,王宁.基于混沌DNA遗传算法的模糊递归神经网络建模[J].控制理论与应用,2011,28(11):1589-1594.
    [5]杨立强,宋海斌,郝天姚.基于BP神经网络的波阻抗反演及应用[J].地球物理学进展,2005,20(1):34-37.
    [6]张新兵,王家林,吴健生.混合最优化算法在地球物理学中的应用现状与前景[J].地球物理学进展,2003,18(2):5-6.

版权所有:© 2023 中国地质图书馆 中国地质调查局地学文献中心