用遗传算法确定三维横向不均匀介质中的近震震源位置

介绍了用遗传算法在三维横向不均匀介质中确定震源位置的研究思路并简述了三维横向不均匀介质中走时计算的基本原理，在遗传算法的迭代过程中，考虑到初始迭代阶段有较大的参数搜索空间而在最后迭代阶段有较小的参数空间，提出了遗传算法中变异概率随迭代次数下降的函数形式并对其进行了分析，结果表明，随迭代次数指数下降的变异概率使得算法收敛加快，而且反演得到的震源位置的各台站走时残差比原定震源位置的明显减小。

HYPOCENTRAL LOCATIONS OF THE NEAR EARTHQUAKES IN 3-D LATERAL HETEROGENEOUS MEDIUM DETERMINED BY USING THE GENETIC ALGORITHMS

The calculation of the travel time in 3 D velocity structure has been given in this paper.In the process of the genetic algorithms,considering large parameter space in the initial iterations and small parameter space in the final iterations,authors give mutation probability which is a large value in the initial iterations and small value in the final iterations.The formation of the mutation probability is linear and exponential with iteration.It is found that the exponential function is a more effective method.Using the exponentially reduced mutation probability,the hypocenter is obtained through genetic algorithms.The residual of the hypocenter is very small.