遗传算法及其在剩余静校正中的应用

遗传算法是近期发展起来的处理非线性优化问题的一种方法。估算剩余静校正量本质上是一个非线性优化问题，而常用于评价解估计的目标函数却是一个具有多极值的非线性问题。当地震资料中的剩余静校正量大，信噪比低时，常规的局部线性反演方法往往易于陷入局部极大值之中，且严重依赖于初始模型的选取。遗传算法则是一种全局搜索方法，能较好地解决这一问题。

Genetic Algorithm and Application to Residual Static Estimation

Genetic algorithm, newly advanced, is used to cope with nonlinear optimization problem. Theestimation of residual statics correction values is essentially a nonlinear inversion problem. The objective functioncommonly used in solution evaluation is nonlinear and multi - extreme. When residual static correction values arehigh and S/ N ratio is low, the conventional inversion methods based on local linearization are usually lost in localmaximum values,and they seriously depend on the selection of initial model. The genetic algorithm is a globalsearch one and is able to solve the problem successfully.