Lazy random walks and optimal transport on graphs (Q726796)

The paper is about the construction of displacement interpolations of probability distributions on a metric graph through the approximation of any optimal transport problem whose cost function is a distance on the graph by a sequence of entropy minimization problems under marginal constraints, called Schrödinger problems, associated with random walks. The displacement interpolations are defined as the limit, in the \(\Gamma\)-convergence set up of the time-marginal flows of the solutions to the relative entropy minimization problems as the jump frequencies of the random walks tend to zero (lazy random walk). The analysis presented in the paper provides also new results about optimal transport on graphs.