Shy couplings (Q863489)

The authors investigate whether shy coupling exists for several classes of Markov processes. A pair \((X,Y)\) of Markov processes is called a Markov coupling if \(X\) and \(Y\) have the same transition probabilities and \((X,Y)\) is a Markov process. A coupling is `shy' if \(\inf_{t\geq 0}\text{dist}(X_t,Y_t)>0\) with positive probability. They focus on two classes of processes: reflected Brownian motions on Euclidean domains and Brownian motions on graphs. They show that there exists a shy coupling for Brownian motions on graphs \(S=E\cup V\) (under some conditions), where \(E\) is the set of edges and \(V\) is the set of vertices. Also, they show that there exist no shy coupling for reflected Brownian motions on \(C^2\)-smooth bounded strictly convex domains.