用遗传算法反演层速度

Interval velocity inversion by genetic algorithm

遗传算法是近期发展起来的处理非线性最优化问题的一种方法。近两年，它已成为地球物理应用领域中的重要研究课题。本文详细介绍了遗传算法的基本原理，并从不同角度研究了遗传算法的效率问题。研究结果表明，遗传算法对处理无井进约束反演、含噪声资料的反演等比较复杂的问题，具有明显的优势。文中对遗传法的稳定性及对噪声的敏感性等问题的讨论说明，遗传算法还有较强的抗干扰能力。作为初步的应用，我们利用遗传算法对实际资料反演了层速度，展示了遗传算法处理实际资料的潜在能力。

Genetic algorithm,newly advanced,is used to cope with nonlinear optimizatinn problem. In recent years,it has become an important subject people research into inapplied geophysics.We detail its principle and analyse its efficiency in different aspects. It is shown that the genetic algorithm very favours the inversion without borehole data restric-tion,and the inversion relating to low signal/noise ratio data. The analyses of its stability and its sensitivity to noise show that the genetic algorithm is quite noise resistant. The genetic algorithm has been initially used in interval velocity inversion of seismic data and displays its potentiality in seismic data processing.