Abstract:The principle of least action is one of the most fundamental physical principle. It says that among all possible motions connecting two points in a phase space, the system will exhibit those motions which extremise an action functional. Many qualitative features of dynamical systems, such as the presence of conservation laws and energy balance equations, are related to the existence of an action functional. Incorporating variational structure into learning algorithms for dynamical systems is, therefore, crucial in order to make sure that the learned model shares important features with the exact physical system. In this paper we show how to incorporate variational principles into trajectory predictions of learned dynamical systems. The novelty of this work is that (1) our technique relies only on discrete position data of observed trajectories. Velocities or conjugate momenta do {\em not} need to be observed or approximated and {\em no} prior knowledge about the form of the variational principle is assumed. Instead, they are recovered using backward error analysis. (2) Moreover, our technique compensates discretisation errors when trajectories are computed from the learned system. This is important when moderate to large step-sizes are used and high accuracy is required. For this, we introduce and rigorously analyse the concept of inverse modified Lagrangians by developing an inverse version of variational backward error analysis. (3) Finally, we introduce a method to perform system identification from position observations only, based on variational backward error analysis.

Dynamically optimal models of atmospheric motion

A variational derivation of the nonequilibrium thermodynamics of a moist atmosphere with rain process and its pseudoincompressible approximation

HAMILTONIAN ALGORITHM FOR SOLVING THE DYNAMIC EQUATIONS OF ATMOSPHERE

Penman and Priestley-Taylor Approaches in Atmospheric Boundary-Layer Dynamics

Entropy and energy conservation for thermal atmospheric dynamics using mixed compatible finite elements

Effective Action for Dissipative and Nonholonomic Systems

Nature's Cost Function: Simulating Physics by Minimizing the Action

New energy, angle momentum and entropy balance approach to modelling climate and macroturbulent atmospheric dynamics, heat and mass transfer at macroscale. III. Low-frequency approximation and singularities in fields of meteoelements

An Unstructured-Mesh Atmospheric Model for Nonhydrostatic Dynamics: Towards Optimal Mesh Resolution

Inertial Manifold of the Atmospheric Equations

Operator Constraint Principle for Simplifying Atmospheric Dynamical Equations

Vortex dynamics: A variational approach using the principle of least action

A fluid particle model

Action principles for dissipative, non-holonomic Newtonian mechanics

The extrema of an action principle for dissipative mechanical systems

Modelling and numerical approximation of a 2.5D set of equations for mesoscale atmospheric processes

A Field-Theory Approach for Modeling Dissipative Relativistic Fluids

Variational Learning of Euler-Lagrange Dynamics from Data

The Hamilton principle for fluid binary mixtures with two temperatures