← Back to arXiv
arXivAnalysis / PDEsarXiv:2608.14904

On the kinetic Fokker--Planck equation in curved geometry

The paper studies how particles move and spread out when they live not in ordinary flat space but on a curved surface or more general curved geometric object called a Riemannian manifold. Think of the difference between describing the random motion of gas molecules on a flat table versus on the surface of a sphere. The governing equation, called the kinetic Fokker-Planck equation, tracks how a probability distribution of particle positions and velocities evolves over time, accounting for both the natural drift of particles along curved paths and random kicks similar to Brownian motion. The paper is written to be largely self-contained, so readers without deep expertise in curved-space mathematics can follow along.

A central technical challenge is that the equation mixes two very different types of variables: position on the curved space and velocity. This mixing makes standard mathematical tools harder to apply. The paper develops and extends a toolkit from functional analysis and spectral theory to handle this situation. Two key phenomena receive careful attention. The first is "hypoellipticity," which means that even though randomness only directly affects the velocity variable, its influence spreads to the position variable over time, making the solution very smooth. The second is "hypocoercivity," which describes how the system relaxes back toward equilibrium at an exponential rate even though the equation is not simply dissipative in an obvious way.

The paper establishes various inequalities that quantify how spread out or uncertain a probability distribution can be, including ones related to spectral gaps and entropy. These tools are applied in two different mathematical settings, one based on squared averages and one based on absolute averages, giving a more complete and unified picture than previous work. The author simplifies and connects several existing results while also proving new ones, and points toward open questions for future research. The work draws motivation from two directions: the physics of relativistic gas dynamics in curved spacetime, and a more abstract mathematical tradition connecting random processes to geometric properties of spaces.

Read original →