Spectral analysis of random walk operators on Euclidean space (Q2428801)

The authors consider random walks on \(\mathbb R^d\) of the following type. For a strictly positive probability density \(\rho\in C^1(\mathbb R^d)\) and \(h>0\), the one-step transition density is given by \(t_h(x,dy)=C\,\rho(y)1_{\{|x-y|<h\}}\,dy\). Let \(T_h\) denote the associated Markov operator. Since, in contrast to related previous work, the state space \(\mathbb R^d\) is not compact, the authors restrict their considerations to densities \(\rho\) being either tempered or Gaussian. For these cases, the spectrum of \(T_h\) (in particular near 1) is analyzed and the results are applied to study convergence of the iterated operator to the stationary measure in terms of total variation distance.