摘要
遗传算法是近年来发展较快的一种求解多参数非线性优化问题的有效方法。本文通过对遗传算法的介绍，对该方法的发展现状及特点进行了分析，在此基础上，对遗传算法进行了改进：通过引入放大系数Ｋ，对目标函数进行压缩及扩展来提高遗传算法的搜索机制；在交换阶段，引入优选机制来体现适者生存这一贯彻遗传算法始终的原则，并把平行遗传算法同参数范围的动态调节结合起来，从而形成了改进的遗传算法。最后用震源和速度结构联合反演的例子来验证其有效性。
Genetic algorithm is a rapid developmental method, which can solve nonlinear multi parameters optimization in recent years. In this paper, by a detailed introduction, the present situation and characteristic of the method is analyzed. Base on this, genetic algorithm is improved. By introduced an amplified parameter, K, objection function can be compressed and amplified in order to improve the search mechanism. In crossover, by introduced the elite selection, the principle of 'survival of the best'' is reflected. Modified genetic algorithm combines parallel genetic algorithms with dynamic adjust of the range of the parameters. At last, the valid of the method is qualified by the example of the simultaneous inversion for velocity structure and focus.
出处
《东北地震研究》
1998年第4期1-14,共14页
Seismological Research of Northeast China
关键词
遗传算法
目标函数
联合反演
速度结构
震源
Genetic algorithm
Objection function
Dynamic adjust
Simultaneous inversion
