A Constrained FEM–LSSVR Framework with B-Spline Basis for Data-Driven Solution of Nonlinear PDEs

Document Type : Research article

Authors

1 Department of Applied Mathematics, Faculty of Mathematical Sciences, Shahid Beheshti University, Tehran, Iran

2 Department of Computer and Data Sciences, Faculty of Mathematical Sciences, Shahid Beheshti University, Tehran, Iran; Institute for Cognitive and Brain Sciences, Shahid Beheshti University, Tehran, Iran

Abstract

This paper presents a constrained hybrid FEM--LSSVR framework for the numerical solution of nonlinear partial differential equations, with a particular focus on FitzHugh--Nagumo (FHN)-type model variants. In the proposed approach, Least Squares Support Vector Regression (LSSVR) is not employed as a post-processing or data-fitting tool; instead, it is formulated as a local primal optimization problem in which the governing PDE, boundary conditions, and inter-element continuity are explicitly enforced as constraints. The finite element method (FEM) is incorporated to provide structural consistency, embedding boundary and continuity information directly into the learning formulation rather than using FEM solutions as training data. The resulting method establishes a tight interaction between numerical discretization and constrained learning, yielding a flexible yet physically consistent approximation framework. The performance of the proposed approach is investigated through numerical experiments on several FHN-type nonlinear PDE variants, including studies with varying mesh resolutions, different final time horizons, and data-driven settings. Comparisons with standard FEM solutions demonstrate improved robustness and accuracy, particularly in scenarios involving limited or noisy data. These results indicate that the constrained FEM--LSSVR formulation provides a viable alternative to traditional numerical solvers for nonlinear PDEs in data-driven contexts.

Keywords

Main Subjects


 

Article PDF

[1] F. Mazhar, C. Bartolucci, F. Regazzoni, M. Paci, L. Dedè, A. Quarteroni, C. Corsi, and S. Severi, A detailed mathematical model of the human atrial cardiomyocyte: Integration of electrophysiology and cardiomechanics, The Journal of Physiology, 602(18), 4543–4583, (2024).
[2] H. Dana Mazraeh, K. Parand, H. Farahani, and S. R. Kheradpisheh, An improved imperialist competitive algorithm for solving an inverse form of the huxley equation, Iranian Journal of Numerical Analysis and Optimization, 14(3), 681–707, (2024).
[3] T. Tanuja, L. Kavitha, S. Varma, V. Raju, C. K. Ganteda, M. Obulesu, W. Jamshed, M. R. Eid, and S. M. Hussain, Heat and mass transfer analysis of casson-based hybrid nanofluid flow in the presence of an aligned magnetic field: An application toward mechanical engineering, Proceedings of the Institution of Mechanical Engineers, Part N: Journal of Nanomaterials, Nanoengineering and Nanosystems, 23977914241248546, (2024).
[4] P. Ferrero, A. Tonini, G. Valenti, M. Chessa, L. Kuthi, P. P. Bassareo, L. Dede, and A. Quarteroni, Appraisal of partial anomalous pulmonary venous drainage through a lumped-parameter mathematical model: a new pathophysiological proof of concept, Interdisciplinary CardioVascular and Thoracic Surgery, 39(5), ivae175, (2024).
[5] T. Lassila and G. Rozza, Parametric free-form shape design with pde models and reduced basis method, Computer Methods in Applied Mechanics and Engineering, 199(23-24), 1583–1592, (2010).
[6] S. W. Teklu, B. B. Terefe, D. K. Mamo, and Y. F. Abebaw, Optimal control strategies on hiv/aids and pneumonia co-infection with mathematical modelling approach, Journal of Biological Dynamics, 18(1), 2288873, (2024).
[7] S. Mohammadi and S. R. Hejazi, Lie symmetry, chaos optimal control in non-linear fractional-order diabetes mellitus, human immunodeficiency virus, migraine parkinsons diseases models: using evolutionary algorithms, Computer Methods in Biomechanics and Biomedical Engineering, 27(5), 651–679, (2024).
[8] T. Eto, H. Garcke, and R. Nürnberg, A parametric finite element method for a degenerate multi-phase Stefan problem with triple junctions, Computational Methods in Applied Mathematics, 26(1), 43–67, (2026).
[9] B. Hooshyarfarzin, M. Abbaszadeh, and M. Dehghan, Numerical simulation and error estimation of the davey-stewartson equations with virtual element method, Applied Mathematics and Computation, 489, 129146, (2025).
[10] K. Soga, Finite difference methods for linear transport equations with sobolev velocity fields, Journal of Mathematical Fluid Mechanics, 27(1), 1–21, (2025).
[11] N. Magome, N. Morita, S. Kaneko, and N. Mitsume, Higher-continuity s-version of finite element method with b-spline functions, 497, Journal of Computational Physics, 497, 112593, (2024).
[12] C.-B. Zhou, Q. Wang, and Y.-X. Ren, Machine learning optimization of compact finite volume methods on unstructured grids, Journal of Computational Physics, 500, 112746, (2024).
[13] M. Reichle, M. Klassen, J. Li, and S. Klinkel, A modified approach for a scaled boundary shell formulation in structural isogeometric analysis, Engineering Analysis with Boundary Elements, 159, 81–94, (2024).
[14] A. Pakniyat, K. Parand, and M. Jani, Least squares support vector regression for differential equations on unbounded domains, Chaos, Solitons & Fractals, 151, 111232, (2021).
[15] C. Pozrikidis, Finite and spectral element methods using matlab, University of California, San Diego, (2005).
[16] R. R. Faria, B. Capron, A. R. Secchi, and M. B. de Souza Jr, A data-driven tracking control framework using physics-informed neural networks and deep reinforcement learning for dynamical systems, Engineering Applications of Artificial Intelligence, 127, 107256, (2024).
[17] M. Raissi, P. Perdikaris, and G. E. Karniadakis, Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations, Journal of Computational Physics,378, 686–707, (2019).
[18] Y. Huang, H. Liu, Y. Zhao, and M. Fei, Chebyshev spectral approximation-based physics-informed neural network for solving higher-order nonlinear differential equations, Engineering with Computers, (2024).
[19] D. Stenkin and V. Gorbachenko, Mathematical modeling on a physics-informed radial basis function network, Mathematics, 12(2), 241, (2024).
[20] M. Movahedian Moghaddam, K. Parand, and S. R. Kheradpisheh, Advanced physics-informed neural network with residuals for solving complex integral equations, Analytical and Numerical Solutions for Nonlinear Equations, 11(1), 153–173, (2026).
[21] S. Mehrkanoon, T. Falck, and J. A. Suykens, Approximate solutions to ordinary differential equations using least squares support vector machines, IEEE Transactions on Neural Networks and Learning Systems, 23(9), 1356–1367, (2012).
[22] K. Parand, A. A. Aghaei, M. Jani, and A. Ghodsi, A new approach to the numerical solution of fredholm integral equations using least squares-support vector regression, Mathematics and Computers in Simulation, 180, 114–128, (2021).
[23] J. A. Rad, K. Parand, and S. Chakraverty, Learning with Fractional Orthogonal Kernel Classifiers in Support Vector Machines: Theory, Algorithms and Applications. Cham, Switzerland: Springer, 2023.
[24] J. Sun, C. Liu, and N. Liu, Data-driven adaptive critic approach for nonlinear optimal control via least squares support vector machine, Asian Journal of Control, 20(1), 104–114, (2018).
[25] M. Bolhassani, H. Dana Mazraeh, and K. Parand, A new method based on least-squares support vector regression for solving optimal control problems, Kybernetika, 60(4), 513–534, (2024).
[26] H. Dana Mazraeh and K. Parand, GELS-SVR: a combination of grammatical evolution and least squares support vector regression for symbolic solution to Falkner–Skan equation, Genetic Programming and Evolvable Machines, 27(1), 3, (2026).
[27] M. Movahedian Moghaddam, H. Dana Mazraeh, and K. Parand, A double deep Q network-powered approach to discovering symbolic solutions for nonlinear integral equations, Engineering Applications of Artificial Intelligence, 169, 114113, (2026).
[28] M. Movahedian Moghaddam, H. Dana Mazraeh, and K. Parand, Symbolic solution of fractional Volterra population models via physics-informed neural networks and distributed particle swarm optimization, Applied Soft Computing, 189, 114512, (2026).
[29] G. E. Karniadakis, I. G. Kevrekidis, L. Lu, P. Perdikaris, S. Wang, and L. Yang, Physics-informed machine learning, Nature Reviews Physics, 3(6), 422–440, (2021).
[30] K. Pelckmans, J. A. Suykens, T. Van Gestel, J. De Brabanter, L. Lukas, B. Hamers, B. De Moor, and J. Vandewalle, Ls-svmlab: a matlab/c toolbox for least squares support vector machines, Tutorial. KULeuven-ESAT. Leuven, Belgium, 142(1-2), (2002).
[31] S. K. Mitusch, S. W. Funke, and M. Kuchta, Hybrid fem-nn models: Combining artificial neural networks with the finite element method, Journal of Computational Physics, 446, 110651, (2021).
[32] J. Cen and Q. Zou, Deep finite volume method for partial differential equations, Journal of Computational Physics 517, 113307, (2024).
[33] M. Asadzadeh, An introduction to the finite element method for differential equations. John Wiley & Sons: John Wiley & Sons, 2020.
[34] K. M. Mohammadi, M. Babaei, Z. Hajimohammadi, and K. Parand, A new numerical method for solving neuro-cognitive models via chebyshev deep neural network (cdnn), Research Square, (2023).
[35] E. Babolian and J. Saeidian, Analytic approximate solutions to burgers, fisher, huxley equations and two combined forms of these equations, Communications in Nonlinear Science and Numerical Simulation, 14(5), 1984–1992, (2009). 
Volume 12, Issue 1
June 2027
Pages 50-67
  • Receive Date: 25 August 2026
  • Revise Date: 08 September 2026
  • Accept Date: 19 September 2026
  • Publish Date: 02 November 2026