PyGpPHs: A Python Package for Bayesian Modeling of Port-Hamiltonian Systems

Authors

Peilun Li, Kaiyuan Tan, Thomas Beckers

Abstract

PyGpPHs is a Python toolbox for physics-informed learning of physical systems. Compared to pure data-driven approaches, it relies on solid physics priors based on the Gaussian process port-Hamiltonian systems (GP-PHS) framework. This foundation guarantees that the learning procedure adheres to the fundamental physical laws governing real-world systems. Utilizing the Bayesian learning method, PyGpPHs enables physics-informed predictions with uncertainty quantification, which are based on the posterior distribution over Hamiltonians. The PyGpPHs toolbox is designed to make Bayesian learning with physics prior accessible to the learning and control community. PyGpPHs can be installed through an open-source link 1.

Keywords

port-Hamiltonian systems; physics-informed learning; Gaussian processes

Citation

• Journal: IFAC-PapersOnLine
• Year: 2024
• Volume: 58
• Issue: 6
• Pages: 54–59
• Publisher: Elsevier BV
• DOI: 10.1016/j.ifacol.2024.08.256
• Note: 8th IFAC Workshop on Lagrangian and Hamiltonian Methods for Nonlinear Control LHMNC 2024- Besançon, France, June 10 – 12, 2024

BibTeX

@article{Li_2024,
  title={{PyGpPHs: A Python Package for Bayesian Modeling of Port-Hamiltonian Systems}},
  volume={58},
  ISSN={2405-8963},
  DOI={10.1016/j.ifacol.2024.08.256},
  number={6},
  journal={IFAC-PapersOnLine},
  publisher={Elsevier BV},
  author={Li, Peilun and Tan, Kaiyuan and Beckers, Thomas},
  year={2024},
  pages={54--59}
}

Download the bib file

References

• Bar-Sinai, Y., Hoyer, S., Hickey, J. & Brenner, M. P. Learning data-driven discretizations for partial differential equations. Proceedings of the National Academy of Sciences vol. 116 15344–15349 (2019) – 10.1073/pnas.1814058116
• Beckers, (2022)
• Cieślik, S. Mathematical Modeling of the Dynamics of Linear Electrical Systems with Parallel Calculations. Energies vol. 14 2930 (2021) – 10.3390/en14102930
• Gardner, (2018)
• Hou, Z.-S. & Wang, Z. From model-based control to data-driven control: Survey, classification and perspective. Information Sciences vol. 235 3–35 (2013) – 10.1016/j.ins.2012.07.014
• Long, (2018)
• Nageshrao, S. P., Lopes, G. A. D., Jeltsema, D. & Babuska, R. Port-Hamiltonian Systems in Adaptive and Learning Control: A Survey. IEEE Transactions on Automatic Control vol. 61 1223–1238 (2016) – 10.1109/tac.2015.2458491
• Rasmussen, (2006)
• Stephany, R. & Earls, C. PDE-READ: Human-readable partial differential equation discovery using deep learning. Neural Networks vol. 154 360–382 (2022) – 10.1016/j.neunet.2022.07.008
• van der Schaft, A. & Jeltsema, D. Port-Hamiltonian Systems Theory: An Introductory Overview. Foundations and Trends® in Systems and Control vol. 1 173–378 (2014) – 10.1561/2600000002
• Wu, K. & Xiu, D. Data-driven deep learning of partial differential equations in modal space. Journal of Computational Physics vol. 408 109307 (2020) – 10.1016/j.jcp.2020.109307
• Zheng, (2022)
• Zobeiry, N. & Humfeld, K. D. A physics-informed machine learning approach for solving heat transfer equation in advanced manufacturing and engineering applications. Engineering Applications of Artificial Intelligence vol. 101 104232 (2021) – 10.1016/j.engappai.2021.104232