Fisher Information Dynamics: A Kinematic Framework for Phase Space Ordering with Applications to Shock Layers and Turbulence
Yingchuan Wu
Abstract
Order here denotes the sharpness of probability-density gradients as quantified by Fisher information, distinct from thermodynamic order parameters. A kinematic framework for phase space ordering is established based on the Fisher information production rate, which possesses a rigorous thermodynamic foundation through the de Bruijn identity, Stam inequality, and dissipation theorem. A four-term decomposition into isotropic contraction, traceless shear, divergence gradient, and boundary flux is derived from the continuity equation. Extending to the Fokker-Planck equation, the non-positivity of the diffusive dissipation term is proved, with strict negativity for non-uniform densities. A vorticity-independence theorem shows that purely rotational velocity fields do not alter the order measure, and explicit geometric sign criteria enable prediction of where order is generated or destroyed. Through exact solutions of the Burgers shock layer, a pointwise identity between compression and divergence-gradient contributions is uncovered, underlying the three-term balance. In two-dimensional incompressible turbulence, a statistical two-term equilibrium (vanishing time-averaged production) is established as the null hypothesis: the conservative reference state from which three-dimensional turbulence departs through vortex stretching. Variational steady states of the associated Wasserstein gradient flow imply vanishing of the total production rate. This framework provides a diagnostic and predictive tool for analyzing ordering dynamics in non-equilibrium systems.
Create a lesson
Related papers
Particle dynamics in a microfluidic T-junction vortex flow: the role of history force
Matteo Chiatto, Luigi de Luca, Howard A. Stone et al.
Predicting the Final Stationary State of Sheared 2D Euler Fluids
Renato Pakter, Yan Levin
Re-emergence of fingering in solidifying liquid-metal drops
Siyu Zhang, Feng Wang, Yuan Ma et al.
The binormal flow: a toy model for turbulence
Elena Danesi, Pablo Merino, Luis Vega
Non intrusive parametric Reduced Order Models based on physics for incompressible flow
Cyrille Allery, Claudine Beghein
Conjugate heat transfer in turbulent pipe flow with azimuthally non-uniform heating
Jonathan Neuhauser, Davide Gatti, Bettina Frohnapfel