PSBP在高密度电阻率法二维反演中的应用.pdfVIP

第 45 卷第6 期 Vo l. 45 No.6 吉林大学学报(工学版) 2015 年 11 月 Journal of Jilin University CEngineering and Technology Edition) Nov. 2015 PSBP 在高密度电阻率法二维反演中的应用 高明亮，于生宝，郑建波，徐畅，张堂，采卉 (吉林大学仪器科学与电气工程学院，长春 130026) 摘 要:针对神经网络算法在非线性反演中容易陷入局部极小、收敛慢、反演精度差等问题，提 出了将粒子群优化(Particle swarm optimization) 算法与 BP 神经网络(Back propagation neu- ral networks) 进行混合反演(简称 PSBP) 。最后，通过经典的地电模型对本文方法的有效性进 行验证，结果表明，本文方法与线性反演方法、BP 神经网络反演方法对比，具有明显的优势，并 取得了很好的反演结果。 关键词:固体地球物理学;粒子群算法;BP 神经网络;高密度电阻率法;反演精度 文章编号:1671-5497(2015)06-2026-08 中圄分类号:P319.3 文献标志码:A DOI: 10. 13229/j. cnki. jdxbgxb201506042 Application of PSBP method in high-density two-dimensional resistivity inversion GAO Ming-liang ,YU Sheng-bao , ZHENG Jian-bo ,XU Chang , ZHANG Kun , LUAN Hui CCollege of Instrumentation and Electrical Engineering , ] ilin University , Changchun 130026 , China) Abstract: In nonlinear inversion , neural network algorithm has the shortcomings of easy to fall into local minimum , slow convergence and poor inversion accuracy. To overcome these shortcomings , the Particle Swarm Optimization (PSO) is combined with Back Propagation (BP) neural network to achieve inversion (referred as PSBP). The effectiveness of the PSBP is verified by the classical geoelectric mode l. Compared with the linear inversion and BP neural network inversion methods , the proposed PSBP method has obvious advantages with good inversion result. Key words: solid geophysics; particle swarm optimization; BP neural networks; high-density resistivity method; inversion accuracy 口 主 目前广泛应用的二维反演软件是由 Loke 编写的 。引