科学技术与应用

基于反向学习策略的萤火虫算法用于瞬变电磁反演

李 朋汕

江西财经大学软件与物联网工程学院

DOI: https://doi.org/10.59429/kxjsyy.v2i3.10709

Keywords: 瞬变电磁反演；萤火虫算法；反向学习；反演

---

Abstract

瞬变电磁反演由于非线性、多解性和病态特性导致传统优化算法（如粒子群算法、差分进化算法），作为 地球物理勘探的关键技术，在局部最优、反演精度不足的情况下，很容易陷入困境。为解决该问题，论文提出基于 萤火虫算法（Firefly Algorithm，FA）与反向学习（Opposition-Based Learning，OBL）的瞬变电磁联合反演方法。该 方法通过反向学习策略优化萤火虫种群初始化，利用反向解的随机性与遍历性提升初始解的多样性，有效拓展全局 搜索空间；同时，结合萤火虫算法的动态吸引度机制与自适应步长策略，平衡算法的全局探索与局部开发能力，降 低早熟收敛概率。

---

References

[1] Raiche A P. Transient electromagnetic field computations for polygonal loops on layered earths[J]. Geophysics,1987(52):785-793.

[2] Ward S H, Hohmann G W. Electromagnetic theory for geophysical exploration Electromagnetic Methods in Applied[J]. Geophysics ed N Nabighian (Tulsa, OK: Society of Exploration Geophysics),1991:121-223.

[3] Xue G Q, Bai C Y, Yan S. Deep sounding TEM investigation method based on a modified fixed central-loop system[J]. Geophys, 2012(76):23-32.

[4] Smith J T, Booker J R. Rapid inversion of two- and threedimensional magnetotelluric data[J]. Journal of Geophysical Research,1991(96):3905-3922.

[5] Constable S C, Parker R L, Constable C G. Occam’s inversion: Apractical algorithm for generating smooth model from electromagnetic sounding data[J]. Geophysics,1987,52(2):289.

[6] Wang G F, Zhao X Y, Ye J C. A regularized conjugate gradient method for wave field transformation of transient electromagnetic field[J]. University politehnica of bucharest scientific bulletin series c-electrical engineering and computer science,2019,81(2)189-202.

[7] Meng C, Han D P, Wang P, et al. Detection of Shallow Strata Using Transient Electromagnetic Method in Case of Mine Goaf[J]. Coal Mine Machinery,2017,38(1):120-122.

[8] Zaslavsky M, Druskin V, Abubakar A, et al. Large-scale GaussNewton inversion of transient controlled-source electromagnetic measurement data using the model reduction framework[J]. Geophysics,2013,78(4):161-171.

[9] Cassiano Antonio Bortolozo, Oleg Bokhonok, Jorge Luís Porsani, et al. Objective function analysis for electric soundings (VES), transient electromagnetic soundings (TEM) and joint inversion VES/TEM[J]. Journal of Applied Geophysics,2017:146.

[10] Irawan D, Grandis H, Sumintadireja P. Quasi-2D resistivity model from inversion of vertical electrical sounding (VES)[J]. Data using guided random search algorithm Journal of Mathematical and Fundamental Sciences,2015,47(3):269-280.

[11] Ogaya X, Ledo J, Queralt P, et al. A layer stripping approach for monitoring resistivity variations using surface magnetotelluric response[J]. Journal of Applied Geophysics,2016(132):100-115.

[12] Egbert G D, Kelbert A. Computational recipes for electromagnetic inverse problems[J]. Geophys. J. Int,2012(189):167-251.

[13] Yang H Y, Li F P, Chen Shen-En,et al. An inversion of transient electromagnetic data from a conical source[J]. Applied Geophysics,2018,15(Z1):545-555+570.

[14] Wei C. Quantum annealing and its application to inversion of acoustic impedance[D]. China University of Geosciences,2015.

[15] Bortolozo C A, Jorge Luís Porsani, Fernando Acácio Monteiro dos Santos, et al. VES/TEM 1D joint inversion by using Controlled Random Search (CRS) algorithm[J]. Journal of Applied Geophysics,2015:112.

[16] Yang X S. Nature-Inspired Metaheuristic Algorithms[M]. Beckington: Luniver Press,2008:79-90.

[17] X-S Yang, S S Sadat Hosseini, A H Gandomi. Firefly algorithm for solving non-convex economic dispatch problems with valve loading effect[J]. Appl. Soft Comput,2012,12(3):1180-1186.

[18] B Basu, G K Mahanti. Fire fly and artificial bees colony algorithm for synthesis of scanned and broadside linear array antenna[J]. Prog. Electromagn. Res. B,2011(32):169-190.

[19] T-F Zhou, G-X Peng, T-Y Hu, et al. Rayleigh wave nonlinear inversion based on the firefly algorithm[J]. Appl. Geophys., 2014,11(2):167-178.

[20] P Wang. Improved firefly algorithm and its application for magnetotelluric sounding data inversion[J]. Prog. Geophys,2019:1-11.

[21] M A Al-Betar, M A Awadallah, A A Heidari, et al. Survival exploration strategies for Harris Hawks Optimizer, Expert Syst[J]. Appl,2021(168).