The nearly analytic discretization of the frequency-domain wave equation produces large-scale, sparse, and ill-conditioned linear system, which challenge conventional iterative solvers. To mitigate this problem, we employ deep neural networks to construct an approximate preconditioner, which is then embedded within BICGSTAB algorithm. This integration retains the convergence characteristics of classical methods while effectively accelerating the iterative process. To be specific, the proposed neural network-based preconditioned solver achieves faster convergence and improved stability compared with classical preconditioned solvers. Numerical tests on representative models show that the method reduces iterative numbers without sacrificing solution accuracy, suggesting its effectiveness for scalable and high-fidelity simulations of frequency-domain wave propagation.
References
[1]
Li, Y., Jia, X., Wu, X. and Geng, Z. (2022) Deep Learning for Enhancing Multisource Reverse Time Migration. IEEE Transactions on Geoscience and Remote Sensing, 60, 1-13. https://doi.org/10.1109/tgrs.2022.3206283
[2]
Moczo, P., Robertsson, J.O.A. and Eisner, L. (2007) The Finite-Difference Time-Domain Method for Modeling of Seismic Wave Propagation. In: Advances in Geophysics, Elsevier, 421-516. https://doi.org/10.1016/s0065-2687(06)48008-0
[3]
Liu, S., Li, X., Wang, W. and Liu, Y. (2014) A Mixed-Grid Finite Element Method with PML Absorbing Boundary Conditions for Seismic Wave Modelling. Journal of Geophysics and Engineering, 11, Article 055009. https://doi.org/10.1088/1742-2132/11/5/055009
[4]
Kosloff, D.D. and Baysal, E. (1982) Forward Modeling by a Fourier Method. Geophysics, 47, 1402-1412. https://doi.org/10.1190/1.1441288
[5]
Komatitsch, D., Tsuboi, S., Tromp, J., Levander, A. and Nolet, G. (2005) The Spectral-Element Method in Seismology. In: Levander, A. and Nolet, G., Eds., Geophysical Monograph Series, American Geophysical Union, 205-227. https://doi.org/10.1029/157gm13
[6]
Yang, D., Lu, M., Wu, R. and Peng, J. (2004) An Optimal Nearly Analytic Discrete Method for 2D Acoustic and Elastic Wave Equations. Bulletin of the Seismological Society of America, 94, 1982-1992. https://doi.org/10.1785/012003155
[7]
Lang, C., Li, Q., Zhou, Y., He, X. and Han, R. (2020) A High-Precision Low-Dispersive Nearly Analytic Difference Method with Its Application in Frequency-Domain Seismic Waveform Inversion. Exploration Geophysics, 51, 355-377. https://doi.org/10.1080/08123985.2019.1699786
[8]
Liu, S., Lang, C., Yang, H. and Wang, W. (2018) A Developed Nearly Analytic Discrete Method for Forward Modeling in the Frequency Domain. Journal of Applied Geophysics, 149, 25-34. https://doi.org/10.1016/j.jappgeo.2017.12.007
[9]
Pratt, R.G. (1999) Seismic Waveform Inversion in the Frequency Domain; Part 1, Theory and Verification in a Physical Scale Model. Geophysics, 64, 888-901. https://doi.org/10.1190/1.1444597
[10]
Plessix, R.-E. (2006) A Review of the Adjoint-State Method for Computing the Gradient of a Functional with Geophysical Applications. Geophysical Journal International, 167, 495-503. https://doi.org/10.1111/j.1365-246x.2006.02978.x
[11]
Erlangga, Y.A., Vuik, C. and Oosterlee, C.W. (2004) On a Class of Preconditioners for Solving the Helmholtz Equation. Applied Numerical Mathematics, 50, 409-425. https://doi.org/10.1016/j.apnum.2004.01.009
[12]
Saad, Y. (2003) Iterative Methods for Sparse Linear Systems. SIAM.
[13]
van der Vorst, H.A. (1992) Bi-CGSTAB: A Fast and Smoothly Converging Variant of Bi-Cg for the Solution of Nonsymmetric Linear Systems. SIAM Journal on Scientific and Statistical Computing, 13, 631-644. https://doi.org/10.1137/0913035
[14]
Saad, Y. and Schultz, M.H. (1986) GMRES: A Generalized Minimal Residual Algorithm for Solving Nonsymmetric Linear Systems. SIAM Journal on Scientific and Statistical Computing, 7, 856-869. https://doi.org/10.1137/0907058
[15]
Du, Z., Liu, J., Liu, W. and Li, C. (2016) Frequency-Space Domain Acoustic Wave Simulation with the BiCGstab (ℓ) Iterative Method. Journal of Geophysics and Engineering, 13, 70-77. https://doi.org/10.1088/1742-2132/13/1/70
[16]
Wang, N. and Lang, C. (2025) Deep Transfer Learning-Based Preconditioned GMRES Method for Acoustic Reverse Time Migration in the Frequency Domain. IEEE Transactions on Geoscience and Remote Sensing, 63, 1-12. https://doi.org/10.1109/tgrs.2025.3556842
[17]
Lang, C. and Ren, Z. (2013) Inexact Rotated Block Triangular Preconditioners for a Class of Block Two-by-Two Matrices. Journal of Engineering Mathematics, 93, 87-98. https://doi.org/10.1007/s10665-013-9674-1
[18]
Varilsuha, D. and Candansayar, M.E. (2018) 3D Magnetotelluric Modeling by Using Finite-Difference Method: Comparison Study of Different Forward Modeling Approaches. Geophysics, 83, WB51-WB60. https://doi.org/10.1190/geo2017-0406.1
[19]
Taghibakhshi, A., Nytko, N., Zaman, T.U., MacLachlan, S., Olson, L. and West, M. (2023) MG-GNN: Multigrid Graph Neural Networks for Learning Multilevel Domain Decomposition Methods. Proceedings of the International Conference on Machine Learning, Honolulu, 23-29 July 2023, 33381-33395.
[20]
Jiang, L., Wang, L., Chu, X., Xiao, Y. and Zhang, H. (2023) PhyGNNet: Solving Spatiotemporal PDEs with Physics-Informed Graph Neural Network. Proceedings of the 2023 2nd Asia Conference on Algorithms, Computing and Machine Learning, Shanghai, 17-19 March 2023, 143-147. https://doi.org/10.1145/3590003.3590029
[21]
Huang, R., Li, R. and Xi, Y. (2022) Learning Optimal Multigrid Smoothers via Neural Networks. SIAM Journal on Scientific Computing, 45, S199-S225. https://doi.org/10.1137/21m1430030
[22]
Lang, C. and Yang, D. (2016) A Nearly Analytic Discrete Method for Solving the Acoustic-Wave Equations in the Frequency Domain. Geophysics, 82, T43-T57. https://doi.org/10.1190/geo2016-0248.1
[23]
Greenbaum, A. (1997) Iterative Methods for Solving Linear Systems. Society for Industrial and Applied Mathematics. https://doi.org/10.1137/1.9781611970937
[24]
Ronneberger, O., Fischer, P. and Brox, T. (2015) U-Net: Convolutional Networks for Biomedical Image Segmentation. In: Navab, N., Hornegger, J., Wells, W. and Frangi, A., Eds., Lecture Notes in Computer Science, Springer International Publishing, 234-241. https://doi.org/10.1007/978-3-319-24574-4_28
[25]
Rahaman, N., Baratin, A., Arpit, D., Draxler, F., Lin, M., Hamprecht, F. and Cour-ville, A. (2019) On the Spectral Bias of Neural Networks. Proceedings of the International Conference on Machine Learning, Long Beach, 9-15 June 2019, 5301-5310.
[26]
Azulay, Y. and Treister, E. (2022) Multigrid-Augmented Deep Learning Preconditioners for the Helmholtz Equation. SIAM Journal on Scientific Computing, 45, S127-S151. https://doi.org/10.1137/21m1433514
[27]
Simonyan, K. and Zisserman, A. (2014) Very Deep Convolutional Networks for Large-Scale Image Recognition. https://arxiv.org/abs/1409.1556
[28]
Lecun, Y., Bottou, L., Bengio, Y. and Haffner, P. (2002) Gradient-Based Learning Applied to Document Recognition. Proceedings of the IEEE, 86, 2278-2324. https://doi.org/10.1109/5.726791
[29]
Dumoulin, V. and Visin, F. (2016) A Guide to Convolution Arithmetic for Deep Learning. https://arxiv.org/abs/1603.07285
[30]
Zhang, H., Cisse, M., Dauphin, Y.N. and Lopez-Paz, D. (2017) Mixup: Beyond Empirical Risk Minimization. https://arxiv.org/abs/1710.09412
[31]
Badrinarayanan, V., Kendall, A. and Cipolla, R. (2017) SegNet: A Deep Convolutional Encoder-Decoder Architecture for Image Segmentation. IEEE Transactions on Pattern Analysis and Machine Intelligence, 39, 2481-2495. https://doi.org/10.1109/tpami.2016.2644615
[32]
He, K., Zhang, X., Ren, S. and Sun, J. (2015) Delving Deep into Rectifiers: Surpassing Human-Level Performance on Imagenet Classification. 2015 IEEE International Conference on Computer Vision (ICCV), Washington, 7-13 December 2015, 1026-1034. https://doi.org/10.1109/iccv.2015.123
[33]
Barron, J.T. (2019) A General and Adaptive Robust Loss Function. 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Long Beach, 15-20 June 2019, 4331-4339. https://doi.org/10.1109/cvpr.2019.00446
[34]
Kingma, D.P. (2014) Adam: A Method for Stochastic Optimization. https://arxiv.org/abs/1412.6980
[35]
Loshchilov, I. and Hutter, F. (2017) Decoupled Weight Decay Regularization. https://arxiv.org/abs/1711.05101
[36]
Lang, C., Liu, S.L., Yang, X.T. and Xu, X.W. (2022) Frequencydomain Acoustic re-Verse Time Migration Based on Improved NAD Method. Chinese Journal of Geo-physics, 65, 1071-1085.
