Abstract
Physics informed neural network (PINN) has become a promising method for solving partial differential equations (PDEs). The loss function of PINN is a weighted sum of multiple items. This makes it easy to fall into local optima, especially the gradient pathologies when solving high frequency problems. The value of penalty coefficients has a crucial impact on the prediction results. Therefore, a new PINN with adaptive penalty coefficients iteratively optimized by biased multiobjective coevolutionary algorithm (BC-PINN) is presented. In BC-PINN, a two-stage optimization mechanism is used to search for parameters of neural network and penalty coefficients respectively. This method involves constructing the fitness function of penalty coefficients based on the biased dominance ranking by data item and regularization item. Compared with the previous works of others, the accuracy of fitting the initial conditions and boundary conditions is considered to be given priority, which is more conducive to PINN converging to the particular solution of PDE. In addition, the set of penalty coefficients is divided into multiple populations to improve the optimization efficiency through coevolutionary algorithm. The empirical results show that: (1) Our method can improve the gradient pathologies and effectively capture the high-frequency features. (2) Compared to the original PINN, it reduces the MSE by 1–6 orders of magnitude in our benchmark functions.
Similar content being viewed by others
Data Availability
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.
References
Strang G, Fix GJ, Griffin DS (1974) An analysis of the finite-element method. J Appl Mech 41:115–126
Ramakrishnan S, Collis SS (2004) Turbulence control simulation using the variational multiscale method. AIAA J 42:745–753
Glatzel T, Litterst C, Cupelli C, Lindemann T, Moosmann C, Niekrawietz R, Streule W, Zengerle R, Koltay P (2008) Computational fluid dynamics (CFD) software tools for microfluidic applications—a case study. Comput Fluids 37(3):218–235
Farmer R, Pike R, Cheng G (2005) CFD analyses of complex flows. Comput Chem Eng 29:2386–2403
Jeong W, Seong J (2014) Comparison of effects on technical variances of computational fluid dynamics (CFD) software based on finite element and finite volume methods. Int J Mech Sci 78:19–26
Dissanayake MWMG, Phan-Thien N (1994) Neural-network-based approximations for solving partial differential equations. Commun Numer Methods Eng 10:195–201
Van Milligen BP, Tribaldos V, Jiménez JA (1995) Neural network differential equation and plasma equilibrium solver. Phys Rev Lett 75:3594
Lagaris IE, Likas A, Fotiadis DI (1998) Artificial neural networks for solving ordinary and partial differential equations. IEEE Trans Neural Networks 9:987–1000
Raissi M, Perdikaris P, Karniadakis GE (2019) Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. J Comput Phys 378:686–707
Raissi M, Yazdani A, Karniadakis GE (2020) Hidden fluid mechanics: Learning velocity and pressure fields from flow visualizations. Science (New York), p 367, 1026–1030.
Ellington EH, Bastille-Rousseau G, Austin C, Landolt KN, Pond BA, Rees EE, Robar N, Murray DL (1981) Automatic differentiation: techniques and applications. J Appl Water Eng Res22:1548–1555
Li R, Wang JX, Lee E, Luo T (2022) Physics-informed deep learning for solving phonon Boltzmann transport equation with large temperature non-equilibrium. Npj Comput Mater, 8.
P.J. Baddoo, B. Herrmann, B.J. McKeon, J. Nathan Kutz, S.L. Brunton, Physics-informed dynamic mode decomposition, Proc. R. Soc. A Math. Phys. Eng. Sci. 479 (2023).
Kunisch K, Walter D (2021) Semiglobal optimal feedback stabilization of autonomous systems via deep neural network approximation. ESAIM Control Optim Calc Var 27:16
Jo H, Son H, Hwang HJ, Kim EH (2020) Deep neural network approach to forward-inverse problems. Networks Heterog Media 15:247–259
Kashefi M, Mukerji T (2022) Physics-informed PointNet: a deep learning solver for steady-state incompressible flows and thermal fields on multiple sets of irregular geometries. J Comput Phys 402:115041
Yuan L, Ni YQ, Deng XY, Hao S (2022) A-PINN: auxiliary physics informed neural networks for forward and inverse problems of nonlinear integro-differential equations. J Comput Phys 462:111260
Xu C, Cao BT, Yuan Y, Meschke G (2022) Transfer learning based physics-informed neural networks for solving inverse problems in tunneling. Comput Methods Appl Mech Eng, 405.
Fan T, Xu K, Pathak J, Darve E. Solving inverse problems in steady state Navier-Stokes equations using deep neural networks, arXiv preprint arXiv:2008.13074v1.
Meng X, Karniadakis GE (2020) A composite neural network that learns from multi-fidelity data: application to function approximation and inverse PDE problems. J Comput Phys 401:109020
Raissi M, Wang Z, Triantafyllou MS, Karniadakis GE (2019) Deep learning of vortex-induced vibrations. J Fluid Mech 861:119–137
Cai S, Wang Z, Fuest F, Jeon YJ, Gray C, Karniadakis GE (2021) Flow over an espresso cup: Inferring 3-D velocity and pressure fields from tomographic background oriented Schlieren via physics-informed neural networks. J Fluid Mech 915:A102
Luo S, Vellakal M, Koric S, Kindratenko V, Cui J (2020) Parameter identification of RANS turbulence model using physics-embedded neural network. Lect Notes Comput Sci 12321:137–149
Eivazi H, Tahani M, Schlatter P, Vinuesa R (2022) Physics-informed neural networks for solving Reynolds-averaged Navier-Stokes equations. Phys Fluids, 34.
Almajid MM, Abu-Al-Saud MO (2022) Prediction of porous media fluid flow using physics informed neural networks. J Pet Sci Eng 208:109205
Haghighat E, Raissi M, Moure A, Gomez H. A deep learning framework for solution and discovery in solid mechanics: linear elasticity, arXiv preprint arXiv:2003.02751.
Rao C, Sun H, Liu Y (2021) Physics-informed deep learning for computational Elastodynamics without labeled data. J Eng Mech 147:04021043
Li W, Bazant MZ, Zhu J (2021) A physics-guided neural network framework for elastic plates: comparison of governing equations-based and energy-based approaches. Comput Methods Appl Mech Eng 383:113933
D.W Abueidda, Q.Lu, S.Koric, Meshless physics-informed deep learning method for three-dimensional solid mechanics, arXiv preprint arXiv:2012.01547v2.
Pu J, Li J, Chen Y (2021) Solving localized wave solutions of the derivative nonlinear Schrödinger equation using an improved PINN method. Nonlinear Dyn 105:1723–1739
Chen Y, Lu L, Karniadakis GE, Dal Negro L (2020) Physics-informed neural networks for inverse problems in nano-optics and metamaterials. Opt Express 28:11618–11633
Pun GPP, Batra R, Ramprasad R, Mishin Y (2019) Physically informed artificial neural networks for atomistic modeling of materials. Nat Commun 10:2339
bin Waheed U, Haghighat E, Alkhalifah T, Song C, Hao Q (2021) PINNeik: Eikonal solution using physics-informed neural networks. Comput Geosci 155:104833.
Wang L, Yan Z (2021) Data-driven rogue waves and parameter discovery in the defocusing nonlinear Schrödinger equation with a potential using the PINN deep learning. Phys Lett Sect A Gen At Solid State Phys 404:127408.
Mishra S, Molinaro R (2021) Physics informed neural networks for simulating radiative transfer. J Quant Spectrosc Radiat Transf 270:107705
Martin CH, Oved A, Chowdhury RA, Ullmann E, Peters NS, Bharath AA. Ep-pinns: cardiac electrophysiology characterisation using physics-informed neural networks. arXiv preprint arXiv:2112.07703.
Patel RG, Manickam I, Trask NA, Wood MA, Lee M, Tomas I, Cyr EC (2022) Thermodynamically consistent physics-informed neural networks for hyperbolic systems. J Comput Phys 449:110754
Lu L, Meng X, Mao Z, Karniadakis GE (2021) DeepXDE: a deep learning library for solving differential equations. SIAM Rev 63:208–228
Peng W, Zhang J, Zhou W, Zhao X, Yao W, Chen X. IDRLnet: a Physics-Informed Neural Network Library. arXiv preprint arXiv:2107.04320.
Chen F, Sondak D, Protopapas P, Mattheakis M, Liu S, Agarwal D, Di Giovanni M (2020) NeuroDiffEq: a python package for solving differential equations with neural networks. J Open Source Softw 5:1931
Yang L, Meng X, Karniadakis GE (2021) B-PINNs: Bayesian physics-informed neural networks for forward and inverse PDE problems with noisy data. J Comput Phys 425:109913
Cheng C, Meng H, Li YZ, Zhang GT (2021) Deep learning based on PINN for solving 2 DOF vortex induced vibration of cylinder. Ocean Eng 240:109932
Fang Z (2021) A high-efficient hybrid physics-informed neural networks based on convolutional neural network. IEEE Trans Neural Netw Learn Syst, pp 1–13.
Gao H, Zahr MJ, Wang JX. Physics-informed graph neural Galerkin networks: aunified framework for solving PDE-governed forward and inverse problems. arXiv preprint arXiv:2107.12146v1.
Arora R. Machine learning-accelerated computational solid mechanics: application to linear elasticity. arXiv preprint arXiv:2112.08676.
Chen X, Duan J, Karniadakis GEM (2021) Learning and meta-learning of stochastic advection-diffusion-reaction systems from sparse measurements. Eur J Appl Math 32:397–420
Wight CL, Zhao J (2021) Solving allen-cahn and cahn-hilliard equations using the adaptive physics informed neural networks, Commun. Comput Phys 29:930–954
Jagtap AD, Kawaguchi K, Karniadakis GE (2020) Adaptive activation functions accelerate convergence in deep and physics-informed neural networks. J Comput Phys 404:109136
Dwivedi V, Parashar N, Srinivasan B (2021) Distributed learning machines for solving forward and inverse problems in partial differential equations. Neurocomputing 420:299–316
Meng X, Li Z, Zhang D, Karniadakis GE (2020) PPINN: Parareal physics-informed neural network for time-dependent PDEs. Comput Methods Appl Mech Eng 370:113250
Jagtap AD, Kharazmi E, Karniadakis GE (2020) Conservative physics-informed neural networks on discrete domains for conservation laws: applications to forward and inverse problems. Comput Methods Appl Mech Eng 365:113028
Moseley B, Markham A, Nissen-Meyer T. Finite Basis Physics-Informed Neural Networks (FBPINNs): a scalable domain decomposition approach for solving differential equations. arXiv preprint arXiv:2107.07871.
Kim Y, Choi Y, Widemann D, Zohdi T (2022) A fast and accurate physics-informed neural network reduced order model with shallow masked autoencoder. J Comput Phys 451:110841
Mishra S, Molinaro R (2022) Estimates on the generalization error of physics-informed neural networks for approximating a class of inverse problems for PDEs. IMA J Numer Anal 42:981–1022
Yu J, Lu L, Meng X, Karniadakis GE (2022) Gradient-enhanced physics-informed neural networks for forward and inverse PDE problems. Comput Methods Appl Mech Eng 393:1–22
Cuomo S, Di Cola VS, Giampaolo F, Rozza G, Raissi M, Piccialli F (2022) Scientific machine learning through physics–informed neural networks: where we are and What’s Next. J Sci Comput, 92.
Wang S, Teng Y, Perdikaris P (2021) Understanding and mitigating gradient pathologies in physics-informed neural networks. SIAM J Sci Comput 43:A3055–A3081
Krishnapriyan S, Gholami A, Zhe S, Kirby Mahoney RM. Characterizing possible failure modes in physics-informed neural networks. arXiv preprint arXiv:2109.01050.
Sun L, Gao H, Pan S, Wang JX (2020) Surrogate modeling for fluid flows based on physics-constrained deep learning without simulation data. Comput Methods Appl Mech Eng 361:112732
Wang S, Yu X, Perdikaris P (2022) When and why PINNs fail to train: a neural tangent kernel perspective. J Comput Phys 449:110768
Xiang Z, Peng W, Zheng X, Zhao X, Yao W. Self-adaptive loss balanced physics-informed neural networks for the incompressible navier-stokes equations. arXiv preprint arXiv:2104.06217.
Cipolla R, Gal Y, Kendall A (2018) Multi-task Learning Using Uncertainty to Weigh Losses for Scene Geometry and Semantics. In: Proc. IEEE Comput. Soc. Conf. Comput. Vis. Pattern Recognit. 2018, pp. 7482–749.
Rall LB (1981) Automatic computation of gradients, Jacobians, Hessians, and applications to optimization. In: Rall LB (ed) Automatic differentiation: techniques and applications. LNCS, vol 120. Springer, Heidelberg, pp 91–111. https://doi.org/10.1007/3-540-10861-0_6
Deb K, Pratap A, Agarwal S, Meyarivan T (2002) A fast and elitist multiobjective genetic algorithm: NSGA-II. IEEE Trans Evol Comput 6:182–197
Li X, Zhang G (2015) Biased multiobjective optimization for constrained single-objective evolutionary optimization. In: Proc. World Congr. Intell. Control Autom, 2015, pp 891–896.
Potter MA, De Jong KA (2000) Cooperative coevolution: an architecture for evolving coadapted subitem. Evol Comput 8:1–29
García-Pedrajas N, Hervás-Martínez C, Ortiz-Boyer D (2005) Cooperative coevolution of artificial neural network ensembles for pattern classification. IEEE Trans Evol Comput 9:271–302
Panait L, Luke S, Wiegand RP (2006) Biasing coevolutionary search for optimal multiagent behaviors. IEEE Trans Evol Comput 10:629–645
Droste S, Jansen T, Wegener I (2002) On the analysis of the (1+1) evolutionary algorithm. Theor Comput Sci 276:51–81
Acknowledgements
The authors appreciate the strong support provided by the National Key R&D Program (2021YFB1714500), the National Ministry Project of China (2019-JCJQ-ZD-049-02, JCKY2019204B021). The authors would like to thank Dr. Liangyue Jia and Dr. Wenbin Ye for their helpful discussions during this work. The authors also thank the anonymous reviewers for their comments and suggestions about improving the manuscript.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Zhu, Z., Hao, J., Huang, J. et al. BC-PINN: an adaptive physics informed neural network based on biased multiobjective coevolutionary algorithm. Neural Comput & Applic 35, 21093–21113 (2023). https://doi.org/10.1007/s00521-023-08876-4
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00521-023-08876-4