A characteristic mapping method for dissipative evolution equations
math.NA, cs.NA, physics.flu-dyn
Submitted: 2026-11-05
Updated: 2026-11-05
License: http://creativecommons.org/licenses/by/4.0/
The gist: The Characteristic Mapping (CM) method is a transport scheme which discretizes the Lagrangian flow map as a composition of spatial interpolants.
Terminology
Abstract
The Characteristic Mapping (CM) method is a transport scheme which discretizes the Lagrangian flow map as a composition of spatial interpolants. This adaptive approximation technique has demonstrated novel resolution properties and the ability to efficiently simulate the fine scales of inviscid turbulent flows. In this work, we construct an extension of this approach to the solution of dissipative problems, including the Navier-Stokes equations. The extension lifts a general scalar evolution equation to a path in a group of diffeomorphisms, where the semigroup approach of the CM method can be exploited. We give an interpretation of this lift through the geometric framework of optimal transport. We then introduce a semidirect product evolution technique to perform the time stepping of the map components. The map is evolved based on the semi-Lagrangian CM approach, and its log-Jacobian is evolved explicitly by an exponential integrator. The numerical scheme is demonstrated to be high-order accurate and is tested on a range of canonical nonlinear dissipative equations. By retaining the resolution properties of the CM method, the approach shows a promising ability to simulate turbulent fluid flows across all Reynolds numbers.
Sources
- A characteristic mapping method for incompressible hydrodynamics on a rotating sphere
- Diffeomorphic Neural Operator Learning
- An Impulse-formed Navier-Stokes Solver based on Long-range Particle Flow Maps
- Universal vector and matrix optimal transport
Related papers
- Do physics-informed neural networks (PINNs) need to be deep? Shallow PINNs using the Levenberg-Marquardt algorithm
- A Neural-preconditioned Poisson Solver for Mixed Dirichlet and Neumann Boundary Conditions
- Advanced Linear Algebra with Applications - Part I (Numerical linear algebra for PDEs, machine learning, and data assimilation)
- Operator learning meets inverse problems: A probabilistic perspective
- Second-order consistency for learning chaotic dynamics via randomized Jacobian matching
- Windowed thinning and query complexity for the bouncy particle and Zigzag samplers