Abstract:The Weak-form Sparse Identification of Nonlinear Dynamics algorithm (WSINDy) has been demonstrated to offer coarse-graining capabilities in the context of interacting particle systems (<a class="link-external link-https" href="https://doi.org/10.1016/j.physd.2022.133406" rel="external noopener nofollow">this https URL</a>). In this work we extend this capability to the problem of coarse-graining Hamiltonian dynamics which possess approximate symmetries associated with timescale separation. Such approximate symmetries often lead to the existence of a Hamiltonian system of reduced dimension that may be used to efficiently capture the dynamics of the symmetry-invariant dependent variables. Deriving such reduced systems, or approximating them numerically, is an ongoing challenge. We demonstrate that WSINDy can successfully identify this reduced Hamiltonian system in the presence of large intrinsic perturbations while remaining robust to extrinsic noise. This is significant in part due to the nontrivial means by which such systems are derived analytically. WSINDy also naturally preserves the Hamiltonian structure by restricting to a trial basis of Hamiltonian vector fields. The methodology is computational efficient, often requiring only a single trajectory to learn the global reduced Hamiltonian, and avoiding forward solves in the learning process. Using nearly-periodic Hamiltonian systems as a prototypical class of systems with approximate symmetries, we show that WSINDy robustly identifies the correct leading-order system, with dimension reduced by at least two, upon observation of the relevant degrees of freedom. We also provide a contribution to averaging theory by proving that first-order averaging at the level of vector fields preserves Hamiltonian structure in nearly-periodic Hamiltonian systems. We provide physically relevant examples, namely coupled oscillator dynamics, the Hénon-Heiles system for stellar motion within a galaxy, and the dynamics of charged particles.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to perform coarse - grained modeling in Hamiltonian systems with approximate symmetries. Specifically, the author extends the capabilities of the WSINDy algorithm to enable it to handle Hamiltonian dynamical systems with approximate symmetries. These systems usually have approximate symmetries due to time - scale separation, and this approximate symmetry can lead to the existence of a low - dimensional Hamiltonian subsystem for efficiently capturing the dynamical behavior of symmetry invariants. ### Background of the Paper 1. **Hamiltonian Systems**: - Hamiltonian systems are a classical mechanics method for describing non - dissipative systems. Hamiltonian equations describe the evolution of points \( z=(q, p) \) in phase space along the level sets of the Hamiltonian function \( H \). - The geometric structure of Hamiltonian systems allows for the systematic exploration of conserved quantities (i.e., constants of motion), which indicate the symmetries of the system. 2. **Approximate Symmetries**: - Approximate symmetry means that in some cases, the symmetry of the system is not strict but approximate. For example, when the system is affected by a small parameter \( \epsilon \), the limit vector field \( X_0 = \lim_{\epsilon \to 0}X_\epsilon \) has exact symmetry. - This approximate symmetry can be used to derive a low - dimensional Hamiltonian subsystem, but this process is very challenging analytically. ### Research Objectives - **Coarse - Grained Modeling**: - The author's goal is to use the WSINDy algorithm to directly identify the low - dimensional subsystem of a Hamiltonian system with approximate symmetries from observational data. - Through this method, the low - dimensional Hamiltonian subsystem can be robustly identified in the presence of large perturbations and external noise. ### Main Contributions 1. **Weak - Form Equation Learning**: - The WSINDy algorithm provides a framework for coarse - graining Hamiltonian systems directly from the observed slow modes by using test functions to interpret the dynamics to be discovered. - This method is computationally efficient, usually only requiring one trajectory to learn the global low - dimensional Hamiltonian system, and avoids the forward - solving process. 2. **Theoretical Results**: - The author proves that the first - order averaged vector field still maintains the Hamiltonian structure in nearly - periodic Hamiltonian systems. This is an important addition to the existing averaging theory. 3. **Application Examples**: - The author provides several physically relevant examples, such as coupled - oscillator dynamics, the Henon - Heiles system (used to describe the motion of stars within a galaxy), and the dynamics of charged particles, demonstrating the effectiveness of the WSINDy algorithm in these systems. ### Conclusion This paper successfully solves the problem of performing coarse - grained modeling in Hamiltonian systems with approximate symmetries by extending the capabilities of the WSINDy algorithm. This method is not only computationally efficient but also shows good robustness in the presence of noise, providing a new tool for understanding and simulating complex physical systems.

Coarse-Graining Hamiltonian Systems Using WSINDy

Coarse-graining Hamiltonian systems using WSINDy

Learning mean-field equations from particle data using WSINDy

Learning Generalized Hamiltonians using fully Symplectic Mappings

Weak SINDy: Galerkin-Based Data-Driven Model Selection

Weak-Form Latent Space Dynamics Identification

Symmetry Preservation in Hamiltonian Systems: Simulation and Learning

Hamiltonian dynamics on the symplectic extended phase space for autonomous and non-autonomous systems

Ensemble WSINDy for Data Driven Discovery of Governing Equations from Laser-based Full-field Measurements

Sparse Symplectically Integrated Neural Networks

Data-Driven Identification of Quadratic Representations for Nonlinear Hamiltonian Systems using Weakly Symplectic Liftings

On Schrödingerization based quantum algorithms for linear dynamical systems with inhomogeneous terms

Simulating the dynamics of large many-body quantum systems with Schrödinger-Feynman techniques

Learning sparse nonlinear dynamics via mixed-integer optimization

Model selection for hybrid dynamical systems via sparse regression

Hamiltonian Matching for Symplectic Neural Integrators

Data-driven reconstruction of chaotic dynamical equations: the Hénon-Heiles type system

Learning of Hamiltonian Dynamics with Reproducing Kernel Hilbert Spaces

Enhancing model identification with SINDy via nullcline reconstruction

Physics-guided weak-form discovery of reduced-order models for trapped ultracold hydrodynamics

Simulating Open Quantum Systems Using Hamiltonian Simulations