A Langevin process reflected at a partially elastic boundary. I (Q655323)

The object of the paper is the Langevin equation \[ \left\{\begin{aligned} & X_t=x+\int_0^t \dot{X}_s\, ds, \\ & \dot{X}_t=u+ B_t-(1+c)\sum_{0<s\leq t} \dot{X}_{s-}\mathbb{I}_{\{ X_s=0\}} +N_t, \end{aligned}\right. \] where \(\mathbb{I}\) denotes the characteristic function, \(B_t\) is the Brownian motion, and \(N_t\) is a continuous nondecreasing process \(N_0=0\) increasing only when the process \((X, \dot{X} )\) hits \((0, 0)\). Denote by \(\operatorname{P}^c_u\) the probability distribution of the vector \((X,\dot{X})\), \(X_0=0\), \(\dot{X}_0=u\). Two main results of the paper are the convergence of the probability laws \(\operatorname{P}^c_u\) of \((X,\dot{X})\), when \(u\to 0+\), and the weak existence and uniqueness of solutions to the above equation with the initial condition \(X_0 = \dot{X}_0 = 0\).