1: \begin{abstract}

2:   Markov chain Monte Carlo (MCMC) is one of the most useful approaches

3:   to scientific computing because of its flexible construction, ease

4:   of use and generality. Indeed, MCMC is indispensable for performing

5:   Bayesian analysis. Two critical

6:   questions that MCMC practitioners need to address are where to start

7:   and when to stop the simulation. Although a great amount of research

8:   has gone into establishing convergence criteria and stopping rules

9:   with sound theoretical foundation, in practice, MCMC users often

10:   decide convergence by applying empirical diagnostic tools. This

11:   review article discusses the most widely used MCMC convergence

12:   diagnostic tools. Some recently proposed stopping rules with

13:   firm theoretical footing are also presented. The convergence diagnostics

14:   and stopping rules are illustrated using three detailed examples.

15: \end{abstract}