参数自适应差分演化算法在面波频散曲线反演中的应用  

Self-adapting control parameters-based differential evolution algorithm for inversion of Rayleigh wave dispersion curves

作　　者：程飞[1] 刘江平[1] 毛茂 王京[1] 宋先海[1] 

机构地区：[1]中国地质大学(武汉)地球内部多尺度成像湖北省重点实验室,湖北武汉430074 [2]中交第二航务工程勘察设计院有限公司,湖北武汉430071

出　　处：《岩土工程学报》2016年第1期147-154,共8页Chinese Journal of Geotechnical Engineering

基　　金：国家自然科学基金项目(41174113)

摘　　要：针对传统差分演化算法在反演瑞雷波频散曲线中存在交叉概率和缩放因子选取不当导致反演结果不准确等问题,在传统差分演化算法的基础上,把控制参数直接编码到个体中,即采用参数自适应差分演化算法对高频瑞雷波频散曲线进行反演来获得近地表的横波速度结构。合成和实际数据的反演实验结果表明:参数自适应差分演化算法继承了传统差分演化算法简单、高效的特点同时,还可在频散曲线的反演中自动地选取合适的参数值并正确地进行反演迭代,无需再通过试验获得交叉概率和缩放因子两个控制参数;频散曲线反演中目标函数的收敛性好,改进算法在迭代的过程中能够快速收敛到全局最优;模型参数的概率分布高,即在大的空间搜索范围内,参数自适应差分演化算法仍然能够准确地搜索到真值范围并找到全局极小值,保证了反演的结果可靠度,使其能有效地应用于瑞雷波频散曲线的反演和解释。The differential evolution algorithm has been widely used in geophysical inversion including inversion of Rayleigh wave dispersion curves. At present the traditional differential evolution algorithm is sensitive to the control parameters set in the process of inversion of Rayleigh wave dispersion curves, and improper selection of the parameters will make the inversion results untrue. Based on the traditional differential evolution algorithm applied in the inversion of Rayleigh wave dispersion curves, the two control parameters, namely crossover probability and zoom factor, are directly coded to individuals, and the differential evolution algorithm with self-adapting control parameters in the inversion of high-frequency Rayleigh wave dispersion curves is adopted to obtain near-surface shear-wave velocity profiles. The results from both synthetic and actual field data demonstrate that：（1） The proposed algorithm not only inherits the simple and efficient features of standard differential evelution algorithm, but also can automatically pick proper parameter values for correct inversion iteration in the inversion of dispersion curves, without relying on the crossover control parameter and amplification factor of the difference vector.（2） The objective function in the proposed algorithm is proved to be able to rapidly converge to the global optimization solution.（3） The wide probability distribution of model parameters, which means the proposed algorithm can define the scope of true-value and find the global minimum even in an extensive search space and guarantee the reliability of inversion results. The proposed algorithm can be applied effectively in the inversion of Rayleigh wave dispersion curves.

关 键 词：瑞雷波 频散曲线反演 差分演化 参数自适应 

分 类 号：TU43[建筑科学—岩土工程]