Multilevel Picard algorithm for general semilinear parabolic PDEs with gradient-dependent nonlinearities
-
Ariel Neufeld
and
Sizhou Wu
Abstract
In this paper we introduce a multilevel Picard approximation algorithm for general semilinear parabolic PDEs with gradient-dependent nonlinearities whose coefficient functions do not need to be constant. We also provide a full convergence and complexity analysis of our algorithm. To obtain our main results, we consider a particular stochastic fixed-point equation (SFPE) motivated by the Feynman–Kac representation and the Bismut–Elworthy–Li formula. We show that the PDE under consideration has a unique viscosity solution which coincides with the first component of the unique solution of the stochastic fixed-point equation. Moreover, the gradient of the unique viscosity solution of the PDE exists and coincides with the second component of the unique solution of the stochastic fixed-point equation. Furthermore, we also provide a numerical example in up to 300 dimensions to demonstrate the practical applicability of our multilevel Picard algorithm.
Funding source: Nanyang Technological University
Award Identifier / Grant number: Nanyang Assistant Professorship Grant (NAP Grant) Machine Learning based Algorithms in Finance and Insurance
-
Research ethics: Not applicable.
-
Informed consent: Not applicable.
-
Author contributions: All authors have accepted responsibility for the entire content of this manuscript and approved its submission.
-
Use of Large Language Models, AI and Machine Learning Tools: None declared.
-
Conflict of interest: The authors state no conflict of interest.
-
Research funding: 1. The Nanyang Assistant Professorship Grant (NAP Grant) Machine Learning based Algorithms in Finance and Insurance. 2. The Fundamental Research Funds for the Central Universities (China).
-
Data availability: The code for the MLP algorithm can be found here: https://github.com/SizhouWu/MLP_Gradient_Nonlinearity.
References
[1] A. Al-Aradi, A. Correia, G. Jardim, D. de Freitas Naiff, and Y. Saporito, “Extensions of the deep Galerkin method,” Appl. Math. Comput., vol. 430, 2022, Art. no. 127287. https://doi.org/10.1016/j.amc.2022.127287.Search in Google Scholar
[2] C. Beck, S. Becker, P. Cheridito, A. Jentzen, and A. Neufeld, “Deep learning based numerical approximation algorithms for stochastic partial differential equations and high-dimensional nonlinear filtering problems,” arXiv preprint arXiv:2012.01194, 2020.Search in Google Scholar
[3] C. Beck, S. Becker, P. Cheridito, A. Jentzen, and A. Neufeld, “Deep splitting method for parabolic PDEs,” SIAM J. Sci. Comput., vol. 43, no. 5, pp. A3135–A3154, 2021. https://doi.org/10.1137/19m1297919.Search in Google Scholar
[4] C. Beck, W. E, and A. Jentzen, “Machine learning approximation algorithms for high-dimensional fully nonlinear partial differential equations and second-order backward stochastic differential equations,” J. Nonlinear Sci., vol. 29, pp. 1563–1619, 2019. https://doi.org/10.1007/s00332-018-9525-3.Search in Google Scholar
[5] C. Beck, M. Hutzenthaler, A. Jentzen, and B. Kuckuck, “An overview on deep learning-based approximation methods for partial differential equations,” arXiv preprint arXiv:2012.12348, 2020.Search in Google Scholar
[6] J. Berner, M. Dablander, and P. Grohs, “Numerically solving parametric families of high-dimensional Kolmogorov partial differential equations via deep learning,” Adv. Neural Inf. Process. Syst., vol. 33, pp. 16 615–16 627, 2020.Search in Google Scholar
[7] J. Castro, “Deep learning schemes for parabolic nonlocal integro-differential equations,” Partial Differ. Equ. Appl., vol. 3, no. 6, p. 77, 2022. https://doi.org/10.1007/s42985-022-00213-z.Search in Google Scholar
[8] P. A. Cioica-Licht, M. Hutzenthaler, and P. T. Werner, “Deep neural networks overcome the curse of dimensionality in the numerical approximation of semilinear partial differential equations,” arXiv preprint arXiv:2205.14398, 2022.Search in Google Scholar
[9] W. E, J. Han, and A. Jentzen, “Deep learning-based numerical methods for high-dimensional parabolic partial differential equations and backward stochastic differential equations,” Commun. Math. Stat., vol. 5, no. 4, pp. 349–380, 2017. https://doi.org/10.1007/s40304-017-0117-6.Search in Google Scholar
[10] W. E and B. Yu, “The deep Ritz method: a deep learning-based numerical algorithm for solving variational problems,” Commun. Math. Stat., vol. 6, no. 1, pp. 1–12, 2018. https://doi.org/10.1007/s40304-018-0127-z.Search in Google Scholar
[11] W. E, J. Han, and A. Jentzen, “Algorithms for solving high dimensional PDEs: from nonlinear Monte Carlo to machine learning,” Nonlinearity, vol. 35, no. 1, pp. 278–310, 2021. https://doi.org/10.1088/1361-6544/ac337f.Search in Google Scholar
[12] R. Frey and V. Köck, “Convergence analysis of the deep splitting scheme: the case of partial integro-differential equations and the associated FBSDEs with jumps,” arXiv preprint arXiv:2206.01597, 2022.Search in Google Scholar
[13] R. Frey and V. Köck, “Deep neural network algorithms for parabolic PIDEs and applications in insurance mathematics,” in Methods and Applications in Fluorescence, Springer, 2022, pp. 272–277.10.1007/978-3-030-99638-3_44Search in Google Scholar
[14] M. Germain, H. Pham, and X. Warin, “Approximation error analysis of some deep backward schemes for nonlinear PDEs,” SIAM J. Sci. Comput., vol. 44, no. 1, pp. A28–A56, 2022. https://doi.org/10.1137/20m1355355.Search in Google Scholar
[15] A. Gnoatto, M. Patacca, and A. Picarelli, “A deep solver for BSDEs with jumps,” arXiv preprint arXiv:2211.04349, 2022.10.2139/ssrn.4271588Search in Google Scholar
[16] L. Gonon, “Random feature neural networks learn Black–Scholes type PDEs without curse of dimensionality,” J. Mach. Learn. Res., vol. 24, no. 189, pp. 1–51, 2023.Search in Google Scholar
[17] L. Gonon and C. Schwab, “Deep ReLU network expression rates for option prices in high-dimensional, exponential Lévy models,” Finance Stoch., vol. 25, no. 4, pp. 615–657, 2021. https://doi.org/10.1007/s00780-021-00462-7.Search in Google Scholar
[18] L. Gonon and C. Schwab, “Deep ReLU neural networks overcome the curse of dimensionality for partial integrodifferential equations,” Anal. Appl., vol. 21, no. 1, pp. 1–47, 2023. https://doi.org/10.1142/s0219530522500129.Search in Google Scholar
[19] P. Grohs, F. Hornung, A. Jentzen, and P. von Wurstemberger, “A proof that artificial neural networks overcome the curse of dimensionality in the numerical approximation of Black–Scholes partial differential equations,” Mem. Am. Math. Soc., vol. 284, 2023. https://doi.org/10.1090/memo/1410.Search in Google Scholar
[20] J. Han, A. Jentzen, and W. E, “Solving high-dimensional partial differential equations using deep learning,” Proc. Natl. Acad. Sci. U. S. A., vol. 115, no. 34, pp. 8505–8510, 2018. https://doi.org/10.1073/pnas.1718942115.Search in Google Scholar PubMed PubMed Central
[21] J. Han and J. Long, “Convergence of the deep BSDE method for coupled FBSDEs,” Probab. Uncertain. Quant. Risk, vol. 5, pp. 1–33, 2020. https://doi.org/10.1186/s41546-020-00047-w.Search in Google Scholar
[22] J. Han, L. Zhang, and E. Weinan, “Solving many-electron Schrödinger equation using deep neural networks,” J. Comput. Phys., vol. 399, 2019, Art. no. 108929.10.1016/j.jcp.2019.108929Search in Google Scholar
[23] C. Huré, H. Pham, and X. Warin, “Deep backward schemes for high-dimensional nonlinear PDEs,” Math. Comput., vol. 89, no. 324, pp. 1547–1579, 2020. https://doi.org/10.1090/mcom/3514.Search in Google Scholar
[24] M. Hutzenthaler, A. Jentzen, T. Kruse, and T. A. Nguyen, “A proof that rectified deep neural networks overcome the curse of dimensionality in the numerical approximation of semilinear heat equations,” SN Partial Differ. Equ. Appl., vol. 1, pp. 1–34, 2020. https://doi.org/10.1007/s42985-019-0006-9.Search in Google Scholar
[25] K. Ito, C. Reisinger, and Y. Zhang, “A neural network-based policy iteration algorithm with global H2-superlinear convergence for stochastic games on domains,” Found. Comput. Math., vol. 21, no. 2, pp. 331–374, 2021. https://doi.org/10.1007/s10208-020-09460-1.Search in Google Scholar
[26] A. Jacquier and M. Oumgari, “Deep curve-dependent PDEs for affine rough volatility,” SIAM J. Financ. Math., vol. 14, no. 2, pp. 353–382, 2023. https://doi.org/10.1137/19m1267805.Search in Google Scholar
[27] A. Jacquier and Z. Zuric, “Random neural networks for rough volatility,” arXiv preprint arXiv:2305.01035, 2023.10.2139/ssrn.4434653Search in Google Scholar
[28] A. Jentzen, D. Salimova, and T. Welti, “A proof that deep artificial neural networks overcome the curse of dimensionality in the numerical approximation of Kolmogorov partial differential equations with constant diffusion and nonlinear drift coefficients,” arXiv preprint arXiv:1809.07321, 2018.Search in Google Scholar
[29] L. Lu, X. Meng, Z. Mao, and G. E. Karniadakis, “DeepXDE: a deep learning library for solving differential equations,” SIAM Rev., vol. 63, no. 1, pp. 208–228, 2021. https://doi.org/10.1137/19m1274067.Search in Google Scholar
[30] J. Y. Nguwi, G. Penent, and N. Privault, “A deep branching solver for fully nonlinear partial differential equations,” arXiv preprint arXiv:2203.03234, 2022.Search in Google Scholar
[31] J. Y. Nguwi, G. Penent, and N. Privault, “Numerical solution of the incompressible Navier–Stokes equation by a deep branching algorithm,” arXiv preprint arXiv:2212.13010, 2022.Search in Google Scholar
[32] J. Y. Nguwi and N. Privault, “A deep learning approach to the probabilistic numerical solution of path-dependent partial differential equations,” Partial Differ. Equ. Appl., vol. 4, no. 4, p. 37, 2023. https://doi.org/10.1007/s42985-023-00255-x.Search in Google Scholar
[33] 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,” J. Comput. Phys., vol. 378, pp. 686–707, 2019. https://doi.org/10.1016/j.jcp.2018.10.045.Search in Google Scholar
[34] C. Reisinger and Y. Zhang, “Rectified deep neural networks overcome the curse of dimensionality for nonsmooth value functions in zero-sum games of nonlinear stiff systems,” Anal. Appl., vol. 18, no. 06, pp. 951–999, 2020. https://doi.org/10.1142/s0219530520500116.Search in Google Scholar
[35] J. Sirignano and K. Spiliopoulos, “DGM: a deep learning algorithm for solving partial differential equations,” J. Comput. Phys., vol. 375, pp. 1339–1364, 2018. https://doi.org/10.1016/j.jcp.2018.08.029.Search in Google Scholar
[36] D. Zhang, L. Guo, and G. E. Karniadakis, “Learning in modal space: solving time-dependent stochastic PDEs using physics-informed neural networks,” SIAM J. Sci. Comput., vol. 42, no. 2, pp. A639–A665, 2020. https://doi.org/10.1137/19m1260141.Search in Google Scholar
[37] C. Beck, L. Gonon, and A. Jentzen, “Overcoming the curse of dimensionality in the numerical approximation of high-dimensional semilinear elliptic partial differential equations,” arXiv preprint arXiv:2003.00596, 2020.Search in Google Scholar
[38] C. Beck, F. Hornung, M. Hutzenthaler, A. Jentzen, and T. Kruse, “Overcoming the curse of dimensionality in the numerical approximation of Allen–Cahn partial differential equations via truncated full-history recursive multilevel Picard approximations,” J. Numer. Math., vol. 28, no. 4, pp. 197–222, 2020. https://doi.org/10.1515/jnma-2019-0074.Search in Google Scholar
[39] S. Becker, R. Braunwarth, M. Hutzenthaler, A. Jentzen, and P. von Wurstemberger, “Numerical simulations for full history recursive multilevel Picard approximations for systems of high-dimensional partial differential equations,” arXiv preprint arXiv:2005.10206, vol. 28, pp. 2109–2138, 2020.10.4208/cicp.OA-2020-0130Search in Google Scholar
[40] W. E, M. Hutzenthaler, A. Jentzen, and T. Kruse, “On multilevel Picard numerical approximations for high-dimensional nonlinear parabolic partial differential equations and high-dimensional nonlinear backward stochastic differential equations,” J. Sci. Comput., vol. 79, no. 3, pp. 1534–1571, 2019. https://doi.org/10.1007/s10915-018-00903-0.Search in Google Scholar
[41] W. E, M. Hutzenthaler, A. Jentzen, and T. Kruse, “Multilevel Picard iterations for solving smooth semilinear parabolic heat equations,” Partial Differ. Equ. Appl., vol. 2, no. 6, pp. 1–31, 2021. https://doi.org/10.1007/s42985-021-00089-5.Search in Google Scholar
[42] M. B. Giles, A. Jentzen, and T. Welti, “Generalised multilevel Picard approximations,” arXiv preprint arXiv:1911.03188, 2019.Search in Google Scholar
[43] M. Hutzenthaler, A. Jentzen, and T. Kruse, “Overcoming the curse of dimensionality in the numerical approximation of parabolic partial differential equations with gradient-dependent nonlinearities,” Found. Comput. Math., vol. 22, no. 4, pp. 905–966, 2022. https://doi.org/10.1007/s10208-021-09514-y.Search in Google Scholar
[44] M. Hutzenthaler, A. Jentzen, T. Kruse, T. Nguyen, and P. von Wurstemberger, “Overcoming the curse of dimensionality in the numerical approximation of semilinear parabolic partial differential equations,” Proc. R. Soc. A, vol. 476, no. 2244, 2020, Art. no. 20190630. https://doi.org/10.1098/rspa.2019.0630.Search in Google Scholar PubMed PubMed Central
[45] M. Hutzenthaler and T. Kruse, “Multilevel Picard approximations of high-dimensional semilinear parabolic differential equations with gradient-dependent nonlinearities,” SIAM J. Numer. Anal., vol. 58, no. 2, pp. 929–961, 2020. https://doi.org/10.1137/17m1157015.Search in Google Scholar
[46] M. Hutzenthaler, A. Jentzen, T. Kruse, and T. A. Nguyen, “Multilevel Picard approximations for high-dimensional semilinear second-order PDEs with Lipschitz nonlinearities,” arXiv preprint arXiv:2009.02484, 2020.Search in Google Scholar
[47] M. Hutzenthaler, A. Jentzen, and P. von Wurstemberger, “Overcoming the curse of dimensionality in the approximative pricing of financial derivatives with default risks,” Electron. J. Probab., vol. 25, no. 101, pp. 1–73, 2020. https://doi.org/10.1214/20-ejp423.Search in Google Scholar
[48] M. Hutzenthaler, T. Kruse, and T. A. Nguyen, “Multilevel Picard approximations for McKean–Vlasov stochastic differential equations,” J. Math. Anal. Appl., vol. 507, no. 1, 2022, Art. no. 125761. https://doi.org/10.1016/j.jmaa.2021.125761.Search in Google Scholar
[49] M. Hutzenthaler and T. A. Nguyen, “Multilevel Picard approximations of high-dimensional semilinear partial differential equations with locally monotone coefficient functions,” Appl. Numer. Math., vol. 181, pp. 151–175, 2022. https://doi.org/10.1016/j.apnum.2022.05.009.Search in Google Scholar
[50] A. Neufeld and S. Wu, “Multilevel Picard approximation algorithm for semilinear partial integro-differential equations and its complexity analysis,” arXiv preprint arXiv:2205.09639, 2022.Search in Google Scholar
[51] E. Fournié, J.-M. Lasry, J. Lebuchoux, P.-L. Lions, and N. Touzi, “Applications of Malliavin calculus to Monte Carlo methods in finance,” Finance Stoch., vol. 3, pp. 391–412, 1999. https://doi.org/10.1007/s007800050068.Search in Google Scholar
[52] E. Gobet, A. Richou, and L. Szpruch, “Numerical approximation of ergodic BSDEs using non linear Feynman–Kac formulas,” arXiv preprint arXiv:2407.09034, 2024.Search in Google Scholar
[53] I. Karatzas and S. Shreve, Brownian Motion and Stochastic Calculus, vol. 113, New York, Springer Science & Business Media, 1991.Search in Google Scholar
[54] X. Mao, Stochastic Differential Equations and Applications, Cambridge, Elsevier, 2007.10.1533/9780857099402Search in Google Scholar
[55] H. Kunita, “Stochastic differential equations based on Lévy processes and stochastic flows of diffeomorphisms,” in Real and Stochastic Analysis, Springer, 2004, pp. 305–373.10.1007/978-1-4612-2054-1_6Search in Google Scholar
[56] A. Neufeld and S. Wu, “Multilevel Picard algorithm for general semilinear parabolic PDEs with gradient-dependent nonlinearities,” arXiv preprint arXiv:2310.12545, 2023.Search in Google Scholar
[57] E. R. Jakobsen and K. H. Karlsen, “Continuous dependence estimates for viscosity solutions of integro-PDEs,” J. Differ. Equ., vol. 212, no. 2, pp. 278–318, 2005. https://doi.org/10.1016/j.jde.2004.06.021.Search in Google Scholar
[58] A. Neufeld, T. A. Nguyen, and S. Wu, “Multilevel Picard approximations overcome the curse of dimensionality in the numerical approximation of general semilinear PDEs with gradient-dependent nonlinearities,” arXiv preprint arXiv:2311.11579, 2023.Search in Google Scholar
[59] G. Barles, R. Buckdahn, and E. Pardoux, “Backward stochastic differential equations and integral-partial differential equations,” Stochastics, vol. 60, nos. 1–2, pp. 57–83, 1997. https://doi.org/10.1080/17442509708834099.Search in Google Scholar
[60] C. Beck, M. Hutzenthaler, and A. Jentzen, “On nonlinear Feynman–Kac formulas for viscosity solutions of semilinear parabolic partial differential equations,” Stoch. Dyn., vol. 21, no. 08, 2021, Art. no. 2150048. https://doi.org/10.1142/s0219493721500489.Search in Google Scholar
© 2025 Walter de Gruyter GmbH, Berlin/Boston
