基于参数自适应的差分进化算法及其在波形反演中的应用
首发时间:2017-02-28
摘要:差分进化算法是一种简单而高效的全局优化算法,在很多实际应用中取到了成功,但其效果依赖于控制参数F和CR的选取。本文提出一种综合考虑个体历史成功信息和邻域个体历史成功信息的控制参数自适应调整策略,并将该策略与CCDE变异策略进行融合,进而应用到波形反演问题中。实验结果表明,在CCDE加入该控制参数自适应调整策略后,算法的收敛速度和收敛精度得到了明显的提升。?????
关键词: 固体地球物理学 差分进化算法 波形反演 参数自适应 CCDE
For information in English, please click here
Differential Evolution with Self-adaptive Control Parameters for High-dimensional Waveform Inversion
Abstract:Differential evolution is a simple and powerful global optimization algorithm, it has applied to many practical problems, but its effect depends on the choice of the control parameters, F and CR. The paper proposes a new method which is gradually self-adaptive by learning from the previous experience of the target individual and neighbor individual. We then embed this self-adaptive method in CCDE and apply it to the waveform inversion. Experiments show that compared to CCDE, the convergence speed of algorithm has a significant improvement and the result of algorithm gets better.
Keywords: Solid Earth Physics differential evolution waveform inversion self-adaptive control parameters CCDE
论文图表:
引用

No.4719550118457214****
同行评议
共计0人参与
勘误表
基于参数自适应的差分进化算法及其在波形反演中的应用
评论
全部评论