免疫算法的改进及在参数反演中的应用

Improvement of Immune Algorithm and its Application in Parameter Inversion

本文提出一种改进的免疫算法。该算法用正交交叉生成初始种群,用精英交叉来增加群体的优良模式,用混合变异提高局部和全局寻优能力。将该方法应用于墨西哥湾地区典型地层模型AVO弹性参数反演。数值试验结果表明,和传统免疫算法相比,改进算法在反演精度和收敛速度上都有了很大的提高。

This paper proposes an improved immune algorithm.This algorithm uses orthogonal crossover to generate initial population and uses elitist-crossover to increase the good patterns of the population and uses hybrid mutation to increase the ability of local and global optimization.This method is applied in AVO inversion for standard elastic models of Mexican gulf.The results show improved algorithm has a higher precision of inversion and a faster speed of convergence than traditional immune algorithm.