1: \begin{abstract}

2: In this paper it is shown that adaptive importance sampling algorithms converge at exponential rate

3: for Markov chain expectation problems that admit a combination of a filtered estimator and a Markov zero-variance measure. 

4: It extends a chain of results---special purpose proofs were already known for several cases~\cite{KBCP99,baggerly2000exponential,desai2001adaptive}.

5: A recent paper~\cite{awad2013zero} provides a complete description of the class of combinations of Markov process expectations of path functionals and 

6: filtered estimators that admit zero-variance importance measures that retain the Markov property. In a way, this is the maximal class for which 

7: adaptive importance sampling algorithms might exhibit exponential convergence.

8: The main purpose of this paper is to prove that this is the case: for (most of) those combinations the natural adaptive importance sampling algorithm converges at exponential rate.

9: In addition, the applicability of general Markov chain theory for this purpose is discussed through the analysis of a counterexample presented in~\cite{desai2001simulation}.

10: \end{abstract}

11: