Consider the biased simple random walk on the circle $\mathbb{Z}/N\mathbb{Z}$, say in continuous time to avoid periodicity where particle hops to left with probability $q$ and right with probability $p$ after a rate one exponential waiting time.

It is easy to see the unique stationary distribution is uniform distribution on the circle and when $p\neq q$, this chain is non-reversible.

My question is: is there any known result about the mixing time for this chain? In particular does it have cutoff? I could not find related results in Levin and Peres' book. Since this chain is non-reversible, I expect the spectral method would not give very good bounds on the mixing time since the eigenvalues are complex...

In general for non-reversible Markov chains to what extend do we know about their mixing times? Is there any good reference for this topic?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.