No label defined (Q4953563)

As usual, by the Markov process the paper understands an operator on \(L_1\) which is a positive contraction (it is supposed that the measure generating \(L_1\) is \(\sigma \)-finite). It is assumed throughout the paper that the integral operator \(K\) is defined by the formula \[ Kf(x) = \int _0^{\lambda (x)}\biggl ( -\dfrac \delta {\delta x}H\bigl (Q\bigl (\lambda (x)\bigr)\bigr) - Q(y)\biggr) f(y) dy, \] where the functions \(Q, \lambda , H\) satisfy simple regularity conditions. It is shown that if \(K\) has no stationary density, then \(K\) is sweeping with respect to the family \(\mathcal A\) of compact subsets of \([0, +\infty)\), i.e., \[ \lim _{n\to \infty} \int _A K^nf dm=0 \] for each set \(A\in \mathcal A\) and every non-negative \(f\) with \(\|f \|=1\).