摘要
针对BP网络和遗传算法(GA)波阻抗反演精度较低且效率不高等问题,提出一种基于BP网络和遗传算法的混合波阻抗反演方法。该方法利用BP算法计算出一定精度的波阻抗初始模型后,再利用改进的遗传算法对该初始模型进行迭代反演,可得到精度更高的反演结果。BP-GA混合波阻抗反演算法简化了二进制的编码解码过程,对BP和GA进行了优化,在保证一定精度的条件下提高了收敛速度,分别用不同子波、不同初始模型和不同噪声试验了该方法的效果,并用实际资料验证了该算法的有效性。
Aimed at the problem of low precision and low efficiency of a- coustic impedance inversion of BP and genetic algorithm (GA) al- gorithm,the seismic impedance combing with the inversion method based on BP neural network and GA is presented.In this method, an initial model with a certain precision of acoustic impedance is cal- culated first by BP neural network and then performed iterative in- version in according to the improved GA algorithm evolution princi- ples.By doing so,a high precision result can be obtained.The method simplifies binary coding and decoding process,optimizes BP and GA algorithm,enhances the convergence rate on condition that a certain precision is ensured.Effect of using this method is proved to be good by different wavelets,initial models,noise experiment, and the actual data processing has also proven the validity of the al- gorithm.
出处
《石油物探》
EI
CSCD
2006年第6期574-579,共6页
Geophysical Prospecting For Petroleum
关键词
BP网络
遗传算法
波阻抗
反演
BP network
genetic algorithm
acoustic impedance
inversion
