An artificial neural network based deep collocation method for the solution of transient linear and nonlinear partial differential equations

Abstract

A combined deep machine learning (DML) and collocation based approach to solve the partial differential equations using artificial neural networks is proposed. The developed method is applied to solve problems governed by the Sine–Gordon equation (SGE), the scalar wave equation and elasto-dynamics. Two methods are studied: one is a space-time formulation and the other is a semi-discrete method based on an implicit Runge–Kutta (RK) time integration. The methodology is implemented using the Tensorflow framework and it is tested on several numerical examples. Based on the results, the relative normalized error was observed to be less than 5% in all cases.

Graphical abstract

Keywords

collocation method / artificial neural networks / deep machine learning / Sine–Gordon equation / transient wave equation / dynamic scalar and elasto-dynamic equation / Runge–Kutta method

References

[1] XuLHuiWZengZ. The algorithm of neural networks on the initial value problems in ordinary differential equations. In: Proceedings of the 2nd IEEE Conference on Industrial Electronics and Applications. New York: Institute of Electrical and Electronics Engineers, 2007,813–816

[2] Mall S, Chakraverty S. Comparison of artificial neural network architecture in solving ordinary differential equations. Advances in Artificial Neural Systems, 2013, 2013: 12–12

[3] YadavNYadavAKumarM. An Introduction to Neural Network Methods for Differential Equations. Berlin: Springer, 2015

[4] Lagaris I E, Likas A, Fotiadis D I. Artificial neural networks for solving ordinary and partial differential equations. IEEE Transactions on Neural Networks, 1998, 9(5): 987–1000

[5] RuddK. Solving partial differential equations using artificial neural networks. Dissertation for the Doctoral Degree. Durham: Duke University, 2013

[6] Al-AradiACorreiaANaiffDJardimGSaporitoY. Solving nonlinear and high-dimensional partial differential equations via deep learning. arXiv, 2018: 1811.08782

[7] Lagaris I E, Likas A C, Papageorgiou D G. Neural-network methods for boundary value problems with irregular boundaries. IEEE Transactions on Neural Networks, 2000, 11(5): 1041–1049

[8] Fuks O, Tchelepi H A. Limitations of physics informed machine learning for nonlinear two-phase transport in porous media. Journal of Machine Learning for Modeling and Computing, 2020, 1(1): 39–61

[9] ChenMMaoSZhangYVictorC MLeungV. Big Data Generation and Acquisition. Springer, 2014, 39–61

[10] Gupta S, Modgil S, Gunasekaran A. Big data in lean six sigma: A review and further research directions. International Journal of Production Research, 2020, 58(3): 947–969

[11] Waheed H, Hassan S U, Aljohani N R, Hardman J, Alelyani S, Nawaz R. Predicting academic performance of students from VLE big data using deep learning models. Computers in Human Behavior, 2020, 104: 106189

[12] Yang R, Yu L, Zhao Y, Yu H, Xu G, Wu Y, Liu Z. Big Data analytics for financial market volatility forecast based on support vector machine. International Journal of Information Management, 2020, 50: 452–462

[13] Thongprayoon C, Kaewput W, Kovvuru K, Hansrivijit P, Kanduri S R, Bathini T, Chewcharat A, Leeaphorn N, Gonzalez-Suarez M L, Cheungpasitporn W. Promises of big data and artificial intelligence in nephrology and transplantation. Journal of Clinical Medicine, 2020, 9(4): 1107

[14] Manzhos S. Machine learning for the solution of the schrodinger equation. Machine Learning: Science and Technology, 2020, 1(1): 013002

[15] Kissas G, Yang Y, Hwuang E, Walter R, Witschey W R, Detre J A, Perdikaris P. Machine learning in cardiovascular flows modeling: Predicting arterial blood pressure from non-invasive 4D flow MRI data using physics-informed neural networks. Computer Methods in Applied Mechanics and Engineering, 2020, 358: 112623

[16] Ghaith M, Li Z. Propagation of parameter uncertainty in swat: A probabilistic forecasting method based on polynomial chaos expansion and machine learning. Journal of Hydrology, 2020, 586: 124854

[17] Russell R D, Shampine L F. A collocation method for boundary value problems. Numerische Mathematik, 1972, 19(1): 1–28

[18] Samaniego E, Anitescu C, Goswami S, Nguyen-Thanh V M, Guo H, Hamdia K, Zhuang X, Rabczuk T. An energy approach to the solution of partial differential equations in computational mechanics via machine learning: Concepts, implementation and applications. Computer Methods in Applied Mechanics and Engineering, 2020, 362: 112790

[19] de Boor C, Swartz B. Collocation at gaussian points. SIAM Journal on Numerical Analysis, 1973, 10(4): 582–606

[20] Raissi M, Perdikaris P, Karniadakis G. Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. Journal of Computational Physics, 2019, 378: 686–707

[21] Goswami S, Anitescu C, Chakraborty S, Rabczuk T. Transfer learning enhanced physics informed neural network for phase-field modeling of fracture. Theoretical and Applied Fracture Mechanics, 2020, 106: 102447

[22] Anitescu C, Atroshchenko E, Alajlan N, Rabczuk T. Artificial neural network methods for the solution of second order boundary value problems. Computers, Materials and Continua, 2019, 59(1): 345–359

[23] Guo H, Zhuang X, Rabczuk T. A deep collocation method for the bending analysis of Kirchhoff plate. Computers, Materials and Continua, 2019, 59(2): 433–456

[24] Nguyen-Thanh V M, Zhuang X, Rabczuk T. A deep energy method for finite deformation hyperelasticity. European Journal of Mechanics. A, Solids, 2020, 80: 103874

[25] Lin J, Zhou S, Guo H. A deep collocation method for heat transfer in porous media: Verification from the finite element method. Journal of Energy Storage, 2020, 28: 101280

[26] Chen A, Zhang X, Zhou Z. Machine learning: Accelerating materials development for energy storage and conversion. InfoMat, 2020, 2(3): 553–576

[27] Hornik K, Stinchcombe M, White H. Multilayer feedforward networks are universal approximators. Neural Networks, 1989, 2(5): 359–366

[28] LuZPuHWangFHuZWangL. The expressive power of neural networks: A view from the width. In: Proceedings of the Conference and Workshop on Neural Information Processing Systems, Advances in neural information processing systems. Cambridge, MA: MIT press, 2017, 6231–6239

[29] Sirignano J, Dgm K S. A deep learning algorithm for solving partial differential equations. Journal of Computational Physics, 2018, 375: 1339–1364

[30] SpeiserD. Discovering the Principles of Mechanics 1600–1800. Berlin: Springer Science & Business Media, 2008

[31] d’AlembertJ R. Tracing back on the curve formed by the rope stretched in vibration. Paris: Memoirs of the Royal Academy of Sciences and Beautiful Literature, 1747 (in French)

[32] Wilson C. D’Alembert versus Euler on the precession of the equinoxes and the mechanics of rigid bodies. Archive for History of Exact Sciences, 2008, 37(3): 233–273

[33] Barone A, Esposito F, Magee C J, Scott A C. Theory and applications of the Sine-Gordon equation. La Rivista del Nuovo Cimento, 1971, 1(2): 227–267

[34] Bour E. Surface deformation theory. Imperial College Magazine, 1862, 19: 1–48

[35] GertE. The Classical Sine-Gordon Equation (SGE). Berlin: Springer, 1981, 93–127

RIGHTS & PERMISSIONS

The Author(s). This article is published with open access at link.springer.com and journal.hep.com.cn