Quantitative spectral gaps and uniform lower bounds in the small noise limit for Markov semigroups generated by hypoelliptic stochastic differential equations

Abstract

We study the convergence rate to equilibrium for a family of Markov semigroups $\{{\mathcal{P}}_t^{\epsilon }{\}}_{\epsilon >0}$ generated by a class of hypoelliptic stochastic differential equations on ${\mathbb{R}}^d$, including Galerkin truncations of the incompressible Navier-Stokes equations, Lorenz-96, and the shell model SABRA. In the regime of vanishing, balanced noise and dissipation, we obtain a sharp (in terms of scaling) quantitative estimate on the exponential convergence in terms of the small parameter $\epsilon$. By scaling, this regime implies corresponding optimal results both for fixed dissipation and large noise limits or fixed noise and vanishing dissipation limits. As part of the proof, and of independent interest, we obtain uniform-in-$\epsilon$ upper and lower bounds on the density of the stationary measure. Upper bounds are obtained by a hypoelliptic Moser iteration, the lower bounds by a de Giorgi-type iteration (both uniform in $\epsilon$). The spectral gap estimate on the semigroup is obtained by a weak Poincaré inequality argument combined with quantitative hypoelliptic regularization of the time-dependent problem.

49 pages

References in corpus (8)

• Spectral gaps in Wasserstein distances and the 2D stochastic Navier--Stokes equations
• Exponential mixing for a class of dissipative PDEs with bounded degenerate noise
• Lower estimates of transition densities and bounds on exponential ergodicity for stochastic PDE's
• Harnack inequality for a class of Kolmogorov-Fokker-Planck equations in non-divergence form
• Moser's estimates for degenerate Kolmogorov equations with non-negative divergence lower order coefficients
• Hölder continuity of solutions to hypoelliptic equations with bounded measurable coefficients
• Ergodicity of Two Dimensional Turbulence
• Lagrangian chaos and scalar advection in stochastic fluid mechanics