【摘要】 遗传算法是近年来发展较快的一种求解非线性优化问题的有效方法。本文通过对遗传算法基本原理的介绍，对该方法的特点进行了分析。我们采用ＷＫＢＪ理论地震图作正演，遗传算法作反演，对用体波波形反演上地幔速度结构的方法进行了研究。在ＷＫＢＪ理论地震图的计算中，通过计算主要射线的平均吸收特征时间，考虑了衰减随距离的变化，通过对不同震源引入不同的虚拟界面，同时对多个地震的波形记录进行反演。探讨了观测误差对反演结果的影响。对理论记录的反演表明，用遗传算法研究上地幔速度结构具有较好的效果。更多还原

【Abstract】 Recently,genetic algorithms is a quickly developed method in solving non-linear optimiza-tion problems。In this paper,the basic principle of genetic algorthms is introduced and themain characteristics is also investigated。Thebody waveform inversion method for upper man-tle velocity structure is studied through genetic algorithms。Formard modeling is achieved bythe means of Chapmen’s WKBJ seismograms。In WKBJ synthetics,the effects of attenuationare considered by calculating the average absorption quality time of the main rays with relative-ly large amplitudes;By introducing different fictitious interface for different sources,the seis-mic records for different earthquakes can be used simultaneously in waveform inversion。Theeffects of different noise levels in observation are also studied。The inversion results for theo-retical records show that the genetic algorthm is effective for the inversion of upper mantle ve-locity structure。更多还原