Abstract:We consider kinetic SDEs with low regularity coefficients in the setting recently introduced in [6]. For the solutions to such equations, we first prove a Harnack inequality. Using the abstract approach of [5], this inequality then allows us to prove, under a Lyapunov condition, the existence and uniqueness (in a suitable class of measures) of a quasi-stationary distribution in cylindrical domains of the phase space. We finally exhibit two settings in which the Lyapunov condition holds: general kinetic SDEs in domains which are bounded in position, and Langevin processes with a non-conservative force and a suitable growth condition on the force.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is to prove the existence and uniqueness of the quasi - stationary distribution (QSD) for kinetic stochastic differential equations (kinetic SDEs) under the condition of having low - regularity coefficients. Specifically, the article mainly focuses on the following issues:
1. **Proof of Harnack inequality**: Prove the Harnack inequality of the solutions of kinetic SDEs under specific low - regularity conditions. The Harnack inequality is an important tool for analyzing partial differential equations and stochastic processes and can provide information on the local behavior of solutions.
2. **Existence and uniqueness of quasi - stationary distribution**: Utilize the above - mentioned Harnack inequality and combine it with the Lyapunov condition to prove the existence and uniqueness of the quasi - stationary distribution of the solutions of kinetic SDEs in a cylindrical domain. The quasi - stationary distribution describes the probability distribution of the system staying within a certain region within a finite time, which is very important for understanding the long - term behavior of the system.
3. **Application of Lyapunov condition**: By constructing an appropriate Lyapunov function, verify that the Lyapunov condition holds in certain cases, thereby ensuring the existence and uniqueness of the quasi - stationary distribution.
### Specific background and motivation
In statistical physics, the evolution of molecular systems at a fixed temperature is usually modeled by the Langevin process:
\[
\begin{cases}
d q_t = M^{-1}p_t \, dt, \\
d p_t = F(q_t) \, dt-\gamma M^{-1}p_t \, dt+\sqrt{2\gamma k T} \, dB_t,
\end{cases}
\]
where \(q_t\) and \(p_t\) represent the position and momentum vectors of particles respectively, \(M\) is the mass matrix, \(\gamma\) is the friction coefficient, \(F\) is the interaction force between particles, \(k\) is the Boltzmann constant, and \(B_t\) is the Brownian motion.
Studying the long - time trajectory sampling of such processes is crucial for understanding the transitions between metastable states of molecular systems, which has wide applications in biology, chemistry, and materials science. However, due to the long time scales required, numerical calculations face challenges. For this reason, many methods rely on the existence of the quasi - stationary distribution of the system's metastable states, that is, the long - time equilibrium reached when the system is trapped in metastable states.
### Main contributions
The main contribution of this paper lies in applying the method in [5] to more general kinetic SDEs, especially in cases with low - regularity coefficients. Specifically:
- Proved the Harnack inequality of the solutions of kinetic SDEs under low - regularity conditions.
- Proved the existence and uniqueness of the quasi - stationary distribution of the solutions of kinetic SDEs in a cylindrical domain when the Lyapunov condition is satisfied.
- Provided two specific settings, showing cases where the Lyapunov condition can be explicitly constructed, including general kinetic SDEs in position - bounded domains, and Langevin processes with non - conservative forces and appropriate growth conditions.
These results provide a theoretical basis for understanding and analyzing the long - term behavior of complex systems.