Использование физически информированных нейронных сетей для численного решения уравнения Блэка–Шоулса
Ключевые слова:
физически информированные нейронные сети
PINN
уравнение Блэка-Шоулса
дифференциальные уравнения в частных производных
Аннотация
Работа посвящена применению физически информированных нейронных сетей (physics-informed neural networks, PINN) для численного решения дифференциальных уравнений в частных производных. Исследования проведены на примере уравнения Блэка–Шоулса. Представлены сравнительные показатели обучения и результатов работы двух сетей. Одна из них является базовой сетью, другая представляет собой модифицированную сеть. Для улучшения работы второй сети использовались инициализация весов Ксавьера, механизм внимания, контроль временн´ой последовательности в обучении, эмбеддинг случайных функций Фурье, адаптивные веса слагаемых функции потерь. Также были использованы разработанные автором ограничения, а именно штраф за отрицательные значения решения уравнения. Результаты показывают, что модификация улучшает качественное поведение решения, особенно в локальной области около точки излома, однако это происходит за счет снижения общей точности и за счет увеличения времени обучения. Так, относительная ошибки базовой PINN после модификации увеличилась, среднее время обучения, приходящееся на одну эпоху, выросло.
Раздел
Методы и алгоритмы вычислительной математики и их приложения
Библиографические ссылки
- M. Malliaris and L. Salchenberger, “A neural network model for estimating option prices,” Applied Intelligence 3 (3), 193–206 (1993).
doi 10.1007/BF00871937
- J. M. Hutchinson, A. W. Lo, and T. Poggio, “A Nonparametric Approach to Pricing and Hedging Derivative Securities via Learning Networks,” The Journal of Finance 49 (3), 851–889 (1994).
doi 10.1111/j.1540-6261.1994.tb00081.x
- F. Black and M. Scholes, “The Pricing of Options and Corporate Liabilities,” Journal of Political Economy 81 (3), 637–654 (1973).
doi 10.1086/260062
- M. Raissi, P. Perdikaris, and G. E. Karniadakis, “Machine learning of linear differential equations using Gaussian processes,” Journal of Computational Physics 348 (11), 683–693 (2017).
doi 10.1016/j.jcp.2017.07.050
- 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).
doi 10.1016/j.jcp.2018.10.045
- A. G. Baydin, B. A. Pearlmutter, A. A. Radul, et al., “Automatic differentiation in machine learning: a survey,” Journal of Machine Learning Research 18 (153), 1–43 (2018).
doi 10.48550/arXiv.1502.05767 Cited August 23, 2026.
- L. B. Rall, Automatic differentiation: techniques and applications(Lecture Notes in Computer Science Vol. 120, 1 ed.) (Springer, Berlin, 1981).
doi 10.1007/3-540-10861-0
- S. Cuomo, V. S. Di Cola, F. Giampaolo, et al., “Scientific Machine Learning through Physics-Informed Neural Networks: Where we are and What’s Next,” Journal of Scientific Computing 92 (3), Article Number 88 (2022).
doi 10.1007/s10915-022-01939-z
- G. E. Karniadakis, I. G. Kevrekidis, L. Lu, et al., “Physics-informed machine learning,” Nature Reviews Physics 3 (6), 422–440 (2021).
doi 10.1038/s42254-021-00314-5
- V. E. Vershinin and R. Ju. Ponomarev, “Application of neural network modeling methods in solving initial boundary value problems for partial differential equations,” Tyumen State University Herald. Physical and Mathematical Modeling. Oil, Gas, Energy 9 (3), 132–147 (2023).
doi 10.21684/2411-7978-2023-9-3-132-147 [in Russian].
- D. Tarkhov, T. Lazovskaya, and V. Antonov, “Adapting PINN Models of Physical Entities to Dynamical Data,” Computation 11 (9), Article Number 168 (2023).
doi 10.3390/computation11090168
- D. P. Zrelova and S. V. Ul’janov, “Physics-Informed Classical Lagrange / Hamilton Neural Networks in Deep Learning,” Modern Information Technologies and IT-Education 18 (2), 310–325 (2022).
doi 10.25559/SITITO.18.202202.310-325 [in Russian].
- T. V. Lazovskaya, D. A. Tarhov, T. A. Shemjakina, and M. D. Choriev, “Evolutionary Approach to PINN Architecture Design for Approximate Solving the Laplace Equation in two Statements: With Discontinuous Initial Condition or with Measurement Data Inside a Square Domain,” Modern Information Technologies and IT-Education 19 (2), 438–446 (2023).
doi 10.25559/sitito.019.202302.438-446 [in Russian].
- F. Gatta, V. S. Di Cola, F. Giampaolo, et al., “Meshless methods for American option pricing through physics-informed neural networks,” Engineering Analysis with Boundary Elements 151 (3), 68–82 (2023).
doi 10.1016/j.enganabound.2023.02.040
- R. Dey, S. N. Srivastava, A. Domoshnitsky, et al., “A data driven study on fractional Black–Scholes-merton equation using physics informed neural network,” Available at SSRN 4995226.
doi 10.2139/ssrn.4995226
- L. Song, Y. Tan, F. Yu, et al., “Optimal approximations for the free boundary problems of the space-time fractional Black–Scholes equations using a combined physics-informed neural network,” Scientific Reports 14 (1), Article Number 25289 (2024).
doi 10.1038/s41598-024-77073-7
- D. Hainaut and A. Casas, “Option pricing in the Heston model with physics inspired neural networks,” Annals of Finance 20 (3), 353–376 (2024).
doi 10.1007/s10436-024-00452-7
- A. Daw, J. Bu, S. Wang, et al., “Mitigating propagation failures in physics-informed neural networks using retain-resample-release (R3) sampling,” in Proceedings of the 40th International Conference on Machine Learning, Honolulu, Hawaii, USA, July 23–29, 2023(PMLR, Vol. 202, 2023), pp. 7264–7302.
https://proceedings.mlr.press/v202/daw23a.html Cited August 23, 2026.
- A. D. Jagtap, K. Kawaguchi, and G. E. Karniadakis, “Adaptive activation functions accelerate convergence in deep and physics-informed neural networks,” Journal of Computational Physics 404 (C), Article Number 109136 (2020).
doi 10.1016/j.jcp.2019.109136
- A. Krishnapriyan, A. Gholami, S. Zhe, et al., “Characterizing possible failure modes in physics-informed neural networks,” in 35th Conference on Neural Information Processing Systems (NeurIPS 2021), Online Conference, December 6–14, 2021(Advances in Neural Information Processing Systems, Vol. 32, 2021), pp. 26548–26560.
https://proceedings.neurips.cc/ Cited August 23, 2026.
- L. D. McClenny and U. M. Braga-Neto, “Self-adaptive physics-informed neural networks,” Journal of Computational Physics 474 (C), Article Number 111722 (2023).
doi 10.1016/j.jcp.2022.111722
- S. Wang, Y. Teng, and P. Perdikaris, “Understanding and Mitigating Gradient Flow Pathologies in Physics-Informed Neural Networks,” SIAM Journal on Scientific Computing 43 (5), A3055–A3081 (2021).
doi 10.1137/20M1318043
- S. Wang, X. Yu, and P. Perdikaris, “When and why PINNs fail to train: A neural tangent kernel perspective,” Journal of Computational Physics 449 (C), Article Number 110768 (2022).
doi 10.1016/j.jcp.2021.110768
- M. M. Dyshaev and V. E. Fedorov, “Symmetries and exact solutions of a nonlinear pricing options equation,” Ufa Mathematical Journal 9 (1), 29–40 (2017).
doi 10.13108/2017-9-1-29
- M. M. Dyshaev and V. E. Fedorov, “Comparing of some sensitivities (Greeks) for nonlinear models of option pricing with market illiquidity,” Mathematical Notes of NEFU 26 (2), 94–108 (2019).
doi 10.25587/SVFU.2019.102.31514
- R. C. Merton, “Theory of rational option pricing,” The Bell Journal of Economics and Management Science 4 (1), 141–183 (1973).
doi 10.2307/3003143
- A. N. Shirjaev, Fundamentals of Stochastic financial Mathematics: Vol. 1 and 2. (MCNMO, M., 2016). [in Russian].
- P. J. Huber, “Robust Estimation of a Location Parameter,” The Annals of Mathematical Statistics 35 (1), 73–101 (1964).
doi 10.1214/aoms/1177703732
- Y. Shin, J. Darbon, and G. E. Karniadakis, “On the convergence of physics informed neural networks for linear second-order elliptic and parabolic type PDEs,” Communications in Computational Physics 28 (5), 2042–2074 (2020).
doi 10.4208/cicp.OA-2020-0193
- D. Liu and Y. Wang, “A Dual-Dimer method for training physics-constrained neural networks with minimax architecture,” Neural Networks 136, 112–125 (2021).
doi 10.1016/j.neunet.2020.12.028
- N. Zobeiry and K. D. Humfeld, “A physics-informed machine learning approach for solving heat transfer equation in advanced manufacturing and engineering applications,” Engineering Applications of Artificial Intelligence 101, Article Number 104232 (2021).
doi 10.1016/j.engappai.2021.104232
- I. A. Chuprov, C. Gao, D. S. Efremenko, et al., “Optimization of Physics-Informed Neural Networks for Nonlinear Schr:odinger Equation,” Doklady RAN. Mat. Inf. Pr. Upr. 514 (2), 28–38 (2023).
doi 10.31857/S2686954323601586 [in Russian].
- X. Glorot and Y. Bengio, “Understanding the difficulty of training deep feedforward neural networks,” Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics (AISTATS), Chia Laguna Resort, Sardinia, Italy, May 13–15, 2010(Proceedings of Machine Learning Research, Vol. 9, 2010), pp. 249–256.
- S.-F. Wang, S. Sankaran, and P. Perdikaris, “Respecting causality for training physics-informed neural networks,” Computer Methods in Applied Mechanics and Engineering 421, Article Number 116813 (2024).
doi 10.1016/j.cma.2024.116813
- R. Mattey and S. Ghosh, “A novel sequential method to train physics informed neural networks for Allen Cahn and Cahn Hilliard equations,” Computer Methods in Applied Mechanics and Engineering 390 (12), Article Number 114474 (2022).
doi 10.1016/j.cma.2021.114474
- A. Vaswani, N. Shazeer, N. Parmar, et al., “Attention is all you need,” in Proceedings of the 31st International Conference on Neural Information Processing Systems (NIPS 2017), Long Beach, CA, USA, December 4–9, 2017Advances in Neural Information Processing Systems. (Curran Associates, Inc., Red Hook, NY, USA, 2017), Vol. 30, pp. 6000–6010.
- Z. Niu, G. Zhong, and H. Yu, “A review on the attention mechanism of deep learning,” Neurocomputing 452, 48–62 (2021).
doi 10.1016/j.neucom.2021.03.091
- N. Rahaman, A. Baratin, D. Arpit, et al., “On the spectral bias of neural networks,” Proceedings of the 36th International Conference on Machine Learning, Long Beach, California, June 9–15, 2019Proceedings of Machine Learning Research (PMLR). (Curran Associates, Inc., Red Hook, NY, USA, 2019), Vol. 97, pp. 5301–5310.
doi 10.48550/arXiv.1806.08734
- Z.-Q. J. Xu, Y. Zhang, and Y. Xiao, “Training behavior of deep neural network in frequency domain,” in Neural Information Processing. 26th International Conference, ICONIP 2019, Sydney, NSW, Australia, December 12–15, 2019Lecture Notes in Computer Science, Vol. 11953. (Springer, Cham., 2019), pp. 264–274.
doi 10.1007/978-3-030-36708-4_22
- Z.-Q. J. Xu, Y. Zhang, T. Luo, et al., “Frequency principle: Fourier analysis sheds light on deep neural networks,” Communications in Computational Physics 28 (5), 1746–1767 (2020).
doi 10.4208/cicp.OA-2020-0085
- M. Tancik, P. Srinivasan, B. Mildenhall, et al., “Fourier features let networks learn high frequency functions in low dimensional domains,” in Proceedings of the 34th Conference on Neural Information Processing Systems (NeurIPS 2020), Vancouver, Canada, December 6–12, 2020Advances in Neural Information Processing Systems. (Curran Associates, Inc., Red Hook, NY, USA, 2020), Vol. 33, pp. 7537–7547.
- A. Jacot, F. Gabriel, and C. Hongler, “Neural tangent kernel: convergence and generalization in neural networks,” in Proceedings of the 32nd Conference on Neural Information Processing Systems (NeurIPS 2018), Montréal, Canada, December 2–8, 2018Advances in Neural Information Processing Systems. (Curran Associates, Inc., Red Hook, NY, USA, 2018), Vol. 31, pp. 8580–8589.
- A. Rahimi and B. Recht, “Random features for large-scale kernel machines,” in Proceedings of the Twenty-First Annual Conference on Neural Information Processing Systems, Vancouver, British Columbia, Canada, December 3-6, 2007Advances in Neural Information Processing Systems (NIPS). (Curran Associates, Inc., Red Hook, NY, USA, 2007), Vol. 20, pp.1177–1184.