A machine learning approach for solving inverse Stefan problem

Document Type : Research Paper

Authors

1 Department of Computer Sciences, Faculty of Mathematical Sciences, Shahid Beheshti University, G.C. Tehran, Iran

2 Department of Statistics and Actuarial Science, University of Waterloo, Waterloo, Canada

Abstract

In this paper, we propose a numerical scheme by using Least Squares Support Vector Regression (LS-SVR) for the simulation of the inverse Stefan problem, which has ill-posedness issues. The purpose of this paper is to express the temperature distribution in a homogeneous environment with a phase change. In the proposed machine learning approach, we apply the unconditionally stable Crank-Nicolson method to decrease the computational cost and reduce one of the dimensions. Therefore, we solve an ODE equation at each time step. The training points of the network are chosen as the Chebyshev roots, which have a normal distribution, and our constructed roots, which we describe more precisely later. In the proposed method, the regularization parameter of the SVM aims to overcome the instability issues, leading to convergent approximation. For the given method, both the primal and dual forms are investigated. The dual form of the problem is written in matrix form. Finally, some numerical examples are provided to illustrate the effectiveness and accuracy of the proposed method.

Keywords

[1] J. Adler and O. Oktem, ¨ Solving ill-posed inverse problems using iterative deep neural networks, Inverse Probl. 33
(2017), no. 12, 124007.
[2] L. Bar and N. Sochen, Unsupervised deep learning algorithm for pde-based forward and inverse problems, 2019.
[3] J. Berg and K. Nystr¨om, Neural network augmented inverse problems for pdes, arXiv, 2018.
[4] J.P. Boyd, Chebyshev and Fourier spectral methods, Courier Corporation, 2001.
[5] P. Dadvand, R. Lopez, and E. Onate, Artificial neural networks for the solution of inverse problems, Proc. Int.
Conf. Design Optim. Meth. Appl. ERCOFTAC, vol. 2006, 2006.
[6] A. Hajiollow, Y. Lotfi, K. Parand, A.H. Hadian, K. Rashedi, and J.A. Rad, Recovering a moving boundary from
Cauchy data in an inverse problem which arises in modeling brain tumor treatment: the (quasi) linearization idea
combined with radial basis functions (RBFs) approximation, Engin. Comput. (2020), 1–15.
[7] J. Ji, C. Zhang, Y. Gui, Q. L¨u, and J. Kodikara, New observations on the application of LS-SVM in slope system
reliability analysis, J. Comput. Civil Engin. 31 (2017), no. 2, 06016002.
[8] M.T. Jin, K.H.and McCann, E. Froustey, and M. Unser, Deep convolutional neural network for inverse problems
in imaging, IEEE Trans. Image Process. 26 (2017), no. 9, 4509–4522.
[9] B.T. Johansson, D. Lesnic, and T. Reeve, A method of fundamental solutions for the one-dimensional inverse
Stefan problem, Appl. Math. Modell. 35 (2011), no. 9, 4367–4378.
[10] A. Karami, S. Abbasbandy, and E. Shivanian, Meshless Local Petrov–Galerkin Formulation of Inverse Stefan
Problem via Moving Least Squares Approximation, Math. Comput. Appl. 24 (2019), no. 4, 101.
[11] Y. Khoo and L. Ying, Switchnet: a neural network model for forward and inverse scattering problems, SIAM J.
Sci. Comput. 41 (2019), no. 5, A3182–A3201.
[12] A. Kirsch, An introduction to the mathematical theory of inverse problems, vol. 120, Springer Science & Business
Media, 2011.
[13] A. Kirsch and A. Rieder, Inverse problems for abstract evolution equations with applications in electrodynamics
and elasticity, Inverse Probl. 32 (2016), no. 8, 085001.
[14] L. Liu, W. Huang, and C. Wang, Hyperspectral image classification with kernel-based least-squares support vector
machines in sum space, IEEE J. Select. Topics Appl. Earth Observ. Remote Sens. 11 (2017), no. 4, 1144–1157.
[15] Y. Lotfi, K. Parand, K. Rashedi, and J. Amani Rad, Numerical study of temperature distribution in an inverse
moving boundary problem using a meshless method, Engin. Comput. (2019), 1–15.
[16] X. Lu, W. Zou, and M. Huang, A novel spatiotemporal LS-SVM method for complex distributed parameter systems
with applications to curing thermal process, IEEE Trans. Ind. Inf. 12 (2016), no. 3, 1156–1165.
[17] A. Lucas, M. Iliadis, R. Molina, and A.K. Katsaggelos, Using deep neural networks for inverse problems in
imaging: beyond analytical methods, IEEE Signal Process. Magazine 35 (2018), no. 1, 20–36.
[18] D. Lv, Q. Zhou, J.K. Choi, J. Li, and X. Zhang, Nonlocal TV-Gaussian prior for Bayesian inverse problems with
applications to limited CT reconstruction, Inverse Probl. Imaging 14 (2020), no. 1, 117.
[19] J. Matlak, D. S lota, and A. Zielonka, Reconstruction of the heat transfer coefficient in the inverse Stefan problem,
Hutnik, Wiadomo´sci Hutnicze 85 (2018), no. 1, 6–9.
[20] S. Mehrkanoon, T. Falck, and J.A. Suykens, Approximate solutions to ordinary differential equations using least
squares support vector machines, IEEE Trans. Neural Networks Learn. Syst. 23 (2012), no. 9, 1356–1367.
[21] S. Mehrkanoon and J.A.K. Suykens, Learning solutions to partial differential equations using LS-SVM, Neurocomput. 159 (2015), 105–116.[22] S. Ozer, C.H. Chen, and H.A. Cirpan, A set of new chebyshev kernel functions for support vector machine pattern
classification, Pattern Recogn. 44 (2011), no. 7, 1435–1447.
[23] L.C. Padierna, M. Carpio, A. Rojas-Dom´─▒nguez, H. Puga, and H. Fraire, A novel formulation of orthogonal
polynomial kernel functions for SVM classifiers: the Gegenbauer family, Pattern Recog. 84 (2018), 211–225.
[24] M. Parhizi and A. Jain, Solution of the Phase Change Stefan Problem With Time-Dependent Heat Flux Using
Perturbation Method, J. Heat Transfer 141 (2019), no. 2, 024503.
[25] L. Pilozzi, F.A. Farrelly, G. Marcucci, and C. Conti, Machine learning inverse problem for topological photonics,
Commun. Phys. 1 (2018), no. 1, 1–7.
[26] J.A. Rad, K. Rashedi, K. Parand, and H. Adibi, The meshfree strong form methods for solving one dimensional
inverse Cauchy-Stefan problem, Engin. Comput. 33 (2017), no. 3, 547–571.
[27] G.M.M. Reddy, M. Vynnycky, and J.A. Cuminato, An efficient adaptive boundary algorithm to reconstruct Neumann boundary data in the MFS for the inverse Stefan problem, J. Comput. Appl. Math. 349 (2019), 21–40.
[28] S. Sarabadan, K. Rashedi, and H. Adibi, Boundary determination of the inverse heat conduction problem in one
and two dimensions via the collocation method based on the satisfier functions, Iran. J. Sci. Technol. Trans. A:
Sci. 42 (2018), no. 2, 827–840.
[29] M.M. Sarsengeldin, S.N.N. Kharin, S. Kassabek, and Z. Mukambetkazin, Exact Solution of the One Phase Inverse
Stefan Problem, Filomat 32 (2018), no. 3, 985–990.
[30] J. Shen, T. Tang, and L.-L. Wang, Spectral methods: algorithms, analysis and applications, vol. 41, Springer
Science & Business Media, 2011.
[31] C.B. Smith and E.M. Hernandez, Non-negative constrained inverse eigenvalue problems–Application to damage
identification, Mech. Syst. Signal Process. 129 (2019), 629–644.
[32] J.A. Suykens, T. Van Gestel, J. De Brabanter, B. De Moor, and J.P. Vandewalle, Least squares support vector
machines, World Scientific, 2002.
[33] J.A.K. Suykens and J. Vandewalle, Least squares support vector machine classifiers, Neural Process. Lett. 9
(1999), no. 3, 293–300.
[34] Y.-H. Wu and H. Shen, Grey-related least squares support vector machine optimization model and its application
in predicting natural gas consumption demand, J. Comput. Appl. Math. 338 (2018), 212–220.
[35] N. Ye, R. Sun, Y. Liu, and L. Cao, Support vector machine with orthogonal chebyshev kernel, 18th Int. Conf.
Pattern Recog. (ICPR’06), vol. 2, IEEE, 2006, pp. 752–755.
[36] N. Zabaras and K. Yuan, Dynamic programming approach to the inverse Stefan design problem, Numerical Heat
Transfer 26 (1994), no. 1, 97–104.
Volume 13, Issue 2
July 2022
Pages 2233-2246
  • Receive Date: 04 June 2020
  • Revise Date: 01 July 2020
  • Accept Date: 07 November 2021