基于遗传算法的叠前地震波形反演构建虚拟井曲线

 鉴于叠前反演分辨率高，叠后反演速度快、稳定性好，所以叠前和叠后混合反演正成为人们研究的热点。其关键技术是叠前地震波形反演构建虚拟井曲线，在此基础上以虚拟井作为控制信息进行叠后反演。本文采用遗传算法（GA）全局寻优的非线性反演方法构建虚拟井曲线，成功地解决了叠前地震波形反演面临的数据和模型之间高度非线性及目标函数具有多个极小值问题。文中给出了利用GA地震波形反演的计算步骤及关键步骤的注意事项。实例结果表明该方法可行。

Using prestack seismic waveform inversion to construct visual curves in wells

In view of the advantages of high resolution of prestack inversion and rapid and stable poststack inversion,the hybrid inversion combining the prestack with poststack becomes bright point that the people vie with each other in studying.The key technique is using prestack seismic waveform inversion to construct visual curves in wells,on that basis,we can do poststack inversion by taking the visual wells as controlling information.The paper used the GA global optimization in non-linear inversion to construct visual curves in wells,successfully solved the issues that highly non-linear between data and model exists in prestack seismic waveform inversion and multiple minimum values in target functions.The paper gave the steps of computation that the seismic waveform inversion is used and matters needing attention in key steps.The cases given in the paper showed the method is feasible.