Semi-classical analysis of a random walk on a manifold (Q2268701): Difference between revisions

The authors study the spectral theory of a Markov chain associated to a natural random walk on a compact connected Riemannian manifold \((M,g)\). The natural random walk is described as a ``ball walk'' on \((M,g)\), where from \(x\), the walk jumps to a uniformly chosen point in a ball of radius \(h\) around the point \(x\). The authors prove a precise Weyl-type estimate on the number of eigenvalues close to \(1\). The second contribution of the authors is the analysis of the associated Metropolis algorithm. The main contribution of this paper is the use of the micro-local analysis as a mean for studying rates of convergence for a Markov chain.