On the spectral gap of Brownian motion with jump boundary (Q638433)

Let \(D \subset \mathbb{R}^d \) be a domain with \(C^{2,\alpha}\)-boundary and let \((\Omega,(B_t)_{t \geq 0}, (\text{P}_x)_{x \in D}) \) be a smooth uniformly elliptic diffusion in \(D\) which is killed after hitting the boundary of \(D\), the diffusion process \((B_t)_{t \geq 0}\) is associated to a generator \(L\) of the form \[ L:\;\frac{1}{2} \sum_{i,j = 1}^d a_{ij}(x)\partial_{ij} + \sum_{i=1}^d b_i(x)\partial _i, \] where the operator is uniformly elliptic with symmetric coefficients. This process induces a compact semigroup of bounded operators \( (P_t^D)_{t \geq 0 }\) in \(L^2(D)\), which is generated by \(L\) with Dirichlet boundary condition, the spectrum consists of the eigenvalues \((\lambda_k^D)_0^\infty\) and the eigenfunctions \((\varphi_k^D)_0^\infty\). Let \((X_t^{\rho,\nu})\) be a diffusion with jump boundary starting from the initial distribution \(\rho\). It is known that this process is uniformly geometrically ergodic, i.e., there exists a probability measure \(\mu^\nu\) such that \[ - \lim_{t \to \infty} \frac{1}{t} \log\sup_{x \in D} \|\text{P}_x(X_t^{\rho,\nu}\epsilon \cdot) - \mu^{\nu}(\cdot)\|_{TV} = \gamma_1(\nu)>0, \] where \(\|\cdot \|_{TV}\) denotes the total variation norm. A first result of this paper is that the spectral gap \(\gamma_1(\nu)\) depends continuously on the jump distribution \(\nu\) with respect to the weak topology. This answers a question posed by \textit{I. Ben-Ari} and \textit{R. G. Pinsky} [J. Funct. Anal. 251, No.~1, 122--140 (2007; Zbl 1137.35045)]. Furthermore, the spectral gap \(\gamma_1(\nu)\) is calculated in the case of the one-dimensional Brownian motion with an arbitrary jump distribution in a direct probabilistic way: \(\gamma_1(\nu) = \lambda_1^D\).