Physics-Informed Neural Networks for PDE-Constrained Optimization and Control

Abstract

The goal of optimal control is to determine a sequence of inputs for maximizing or minimizing a given performance criterion subject to the dynamics and constraints of the system under observation. This work introduces Control Physics-Informed Neural Networks (PINNs), which simultaneously learn both the system states and the optimal control signal in a single-stage framework that leverages the system’s underlying physical laws. While prior approaches often follow a two-stage process-modeling, the system first and then devising its control—the presented novel framework embeds the necessary optimality conditions directly into the network architecture and loss function. We demonstrate the effectiveness of the novel methodology by solving various open-loop optimal control problems governed by analytical, one-dimensional, and two-dimensional partial differential equations (PDEs).

Keywords

Optimal control / Scientific machine learning (SciML) / Physics-informed neural networks (PINNs) / 65K05 / 35Q93 / 90C46

References

[1] Antonelo, E.A., Camponogara, E., Seman, L.O., Jordanou, J.P., de Souza, E.R., Hübner, J.F.: Physics-informed neural nets for control of dynamical systems. Neurocomputing 579, 127419 (2024)

[2] Atzberger, P. J. Importance of the mathematical foundations of machine learning methods for scientific and engineering applications. arXiv:1808.02213 (2018)

[3] Bellman, R.: Dynamic programming. Science 153(3731), 34–37 (1966). (Accessed 2022-05-15)

[4] Biegler, L. T., Ghattas, O., Heinkenschloss, M., van Bloemen Waanders, B. Large-scale PDE-constrained optimization: an introduction. In: Lecture Notes in Computational Science and Engineering, vol. 30. Springer, Berlin/Heidelberg (2003)

[5] Blonigan, P.J., Fernandez, P., Murman, S.M., Wang, Q., Rigas, G., Magri, L.: Toward a chaotic adjoint for LES (2017). https://doi.org/10.48550/ARXIV.1702.06809

[6] ChenY, LuL, KarniadakisGE, Dal NegroL. Physics-informed neural networks for inverse problems in nano-optics and metamaterials. Opt. Express, 2020, 28811618.

[7] Chen, Y., Shi, Y., Zhang, B.: Optimal control via neural networks: a convex approach. arXiv:1805.11835 (2018)

[8] Chennault, A., Popov, A.A., Subrahmanya, A.N., Cooper, R., Karpatne, A., Sandu, A.: Adjoint-matching neural network surrogates for fast 4D-Var data assimilation (2021). https://doi.org/10.48550/ARXIV.2111.08626

[9] CuomoS, Di ColaVS, GiampaoloF, RozzaG, RaissiM, PiccialliF. Scientific machine learning through physics-informed neural networks: where we are and what’s next. J. Sci. Comput., 2022, 92388

[10] Daw, A., Bu, J., Wang, S., Perdikaris, P., Karpatne, A.: Rethinking the importance of sampling in physics-informed neural networks (2022). https://doi.org/10.48550/ARXIV.2207.02338

[11] De FlorioM, SchiassiE, CalabròF, FurfaroR. Physics-informed neural networks for 2nd order odes with sharp gradients. J. Comput. Appl. Math., 2024, 436115396

[12] GokhaleG, ClaessensB, DevelderC. Physics informed neural networks for control oriented thermal modeling of buildings. Appl. Energy, 2022, 314. 118852

[13] HagerWW. Runge-Kutta methods in optimal control and the transformed adjoint system. Numer. Math., 2000, 87: 247-282

[14] HairerE, WannerGSolving Ordinary Differential Equations II: Stiff and Differential-Algebraic Problems, 2002, Berlin. Springer.

[15] HeQ, Barajas-SolanoD, TartakovskyG, TartakovskyAM. Physics-informed neural networks for multiphysics data assimilation with application to subsurface transport. Adv. Water Resour., 2020, 141. 103610

[16] Hinze, M., Pinnau, R., Ulbrich, M., Ulbrich, S.: Optimization with PDE Constraints. Springer, Dordrecht, Netherlands (2008). https://doi.org/10.1007/978-1-4020-8839-1

[17] HornikK, StinchcombeM, WhiteH. Multilayer feedforward networks are universal approximators. Neural Netw., 1989, 2(5): 359-366

[18] HwangR, LeeJY, ShinJY, HwangHJ. Solving PDE-constrained control problems using operator learning. Proceedings of the AAAI Conference on Artificial Intelligence, 2022, 36(4): 4504-4512

[19] Kollmannsberger, S., D’Angella, D., Jokeit, M., Herrmann, L.: Physics-Informed Neural Networks, pp. 55–84. Springer, Cham (2021). https://doi.org/10.1007/978-3-030-76587-3-5

[20] Krishnapriyan, A.S., Gholami, A., Zhe, S., Kirby, R.M., Mahoney, M.W.: Characterizing possible failure modes in physics-informed neural networks (2021). https://doi.org/10.48550/ARXIV.2109.01050

[21] LotkaAJ. Contribution to the theory of periodic reactions. J. Phys. Chem., 1910, 14(3): 271-274.

[22] LuL, MengX, MaoZ, KarniadakisG. DeepXDE: a deep learning library for solving differential equations. SIAM Rev., 2021, 63(1): 208-228.

[23] Lu, L., Pestourie, R., Yao, W., Wang, Z., Verdugo, F., Johnson, S.G.: Physics-informed neural networks with hard constraints for inverse design (2021). https://doi.org/10.48550/ARXIV.2102.04626

[24] MowlaviS, NabiS. Optimal control of PDEs using physics-informed neural networks. J. Comput. Phy., 2023, 473111731

[25] MropeF, KigodiO. Modeling the transmission dynamics of maize foliar disease in maize plants: modeling maize foliar disease dynamics in maize plants. J. Math. Anal. Model., 2024, 5(2): 114-135

[26] Mrope, F., Nyerere, N.: Modeling the transmission dynamics of bushfires on cashew nut production. J. Math. Anal. Model. 5(2), 98–113 (2024). https://doi.org/10.48185/jmam.v5i2.1183

[27] NabianMA, MeidaniH. A deep learning solution approach for high-dimensional random differential equations. Probab. Eng. Mech., 2019, 57: 14-25.

[28] Nellikkath, R., Chatzivasileiadis, S.: Physics-informed neural networks for minimising worst-case violations in DC optimal power flow. In: IEEE International Conference on Communications, Control, and Computing Technologies for Smart Grids (SmartGridComm), pp. 419–424 (2021)

[29] Nicodemus, J., Kneifl, J., Fehr, J., Unger, B.: Physics-informed neural networks-based model predictive control for multi-link manipulators (2021). https://doi.org/10.48550/ARXIV.2109.10793

[30] NocedalJ, WrightSNumerical Optimization, 1999, New York. Springer.

[31] PangG, LuL, KarniadakisGE. fPINNs: fractional physics-informed neural networks. SIAM J. Sci. Comput., 2019, 41(4): 2603-2626.

[32] Raissi, M., Perdikaris, P., Karniadakis, G.E.: Physics informed deep learning (part I): data-driven solutions of nonlinear partial differential equations. arXiv:1711.10561 (2017a)

[33] Raissi, M., Perdikaris, P., Karniadakis, G.E.: Physics informed deep learning (part II): data-driven discovery of nonlinear partial differential equations. arXiv:1711.10566 (2017b)

[34] RaissiM, PerdikarisP, KarniadakisGE. Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. J. Comput. Phys., 2019, 378: 686-707

[35] RaissiM, YazdaniA, KarniadakisG. Hidden fluid mechanics: learning velocity and pressure fields from flow visualizations. Science, 2020, 3674741.

[36] Roberts, S., Popov, A.A., Sandu, A.: ODE test problems: a MATLAB suite of initial value problems. arXiv:1901.04098 (2019)

[37] RobertsS, PopovAA, SarsharA, SanduA. A fast time-stepping strategy for dynamical systems equipped with a surrogate model. SIAM J. Sci. Comput., 2022, 44(3): 1405-1427.

[38] Shafiullah, K.: Existence theory and stability analysis to a class of hybrid differential equations using confirmable fractal fractional derivative. J. Fract. Calc. Nonlinear Syst. 7(8) (2024)

[39] ShahK, AbdeljawadT, AlrabaiahH. On coupled system of drug therapy via piecewise equations. Fractals, 2022, 30082240206

[40] ShahK, SarwarM, AbdeljawadT. On mathematical model of infectious disease by using fractals fractional analysis. Discr. Cont. Dyn. Syst., 2024, 17(10): 3064-3085

[41] ShahK, SinanM, AbdeljawadT, El-ShorbagyMA, AbdallaB, AbualrubMS. A detailed study of a fractal-fractional transmission dynamical model of viral infectious disease with vaccination. Complexity, 2022, 202217236824

[42] Son, H., Cho, S.W., Hwang, H.J.: AL-PINNs: augmented Lagrangian relaxation method for physics-informed neural networks (2022). https://doi.org/10.48550/ARXIV.2205.01059

[43] Wang, S., Bhouri, M.A., Perdikaris, P.: Fast PDE-constrained optimization via self-supervised operator learning. arXiv:2110.13297 (2021)

[44] Wang, S., Teng, Y., Perdikaris, P.: Understanding and mitigating gradient pathologies in physics-informed neural networks (2020). https://doi.org/10.48550/ARXIV.2001.04536

[45] Wang, S., Wang, H., Perdikaris, P.: Improved architectures and training algorithms for deep operator networks (2021). https://doi.org/10.48550/ARXIV.2110.01654

[46] Willard, J., Jia, X., Xu, S., Steinbach, M., Kumar, V.: Integrating scientific knowledge with machine learning for engineering and environmental systems ACM Comput. Surv. 55(4), 1–37, 66 (2023)

[47] YanX-B, XuZ-QJ, MaZ. Bayesian Inversion with Neural Operator (BINO) for modeling subdiffusion: forward and inverse problems. J. Comput. Appl. Math., 2025, 454. 116191

[48] YangL, ZhangD, KarniadakisGE. Physics-informed generative adversarial networks for stochastic differential equations. SIAM J. Sci. Comp., 2020, 42(1): A292-A317

[49] YuTF, ChungET, CheungKC, ZhaoL. Learning computational upscaling models for a class of convection-diffusion equations. J. Comput. Appl. Math., 2024, 445. 115814

[50] ZhangD, GuoL, KarniadakisGE. Learning in modal space: solving time-dependent stochastic PDEs using physics-informed neural networks. SIAM J. Sci. Comp., 2020, 42(2): A639-A665

[51] ZhangD, LuL, GuoL, KarniadakisGE. Quantifying total uncertainty in physics-informed neural networks for solving forward and inverse stochastic problems. J. Comput. Phys., 2019, 397. 108850

Funding

U.S. Department of Energy(ASCR DE-SC0021313)

National Science Foundation(DMS-2411069)

RIGHTS & PERMISSIONS

The Author(s)

Just Accepted

This article has successfully passed peer review and final editorial review, and will soon enter typesetting, proofreading and other publishing processes. The currently displayed version is the accepted final manuscript. The officially published version will be updated with format, DOI and citation information upon launch. We recommend that you pay attention to subsequent journal notifications and preferentially cite the officially published version. Thank you for your support and cooperation.