1: \begin{abstract}

2: \noindent {\rm

3: Adaptive Multilevel Splitting (AMS for short) is a generic Monte Carlo method for Markov processes that simulates rare events and estimates associated probabilities. Despite its practical efficiency, there are almost no theoretical results on the convergence of this algorithm. The purpose of this paper is to prove both consistency and asymptotic normality results in a general setting. This is done by associating to the original Markov process a level-indexed process, also called a stochastic wave, and by showing that AMS can then be seen as a Fleming-Viot type particle system. This being done, we can finally apply general results on Fleming-Viot particle systems that we have recently obtained.

4: %The Adaptive Multilevel Splitting, in the last particle version, is a rare event particle simulation method working as follows. $N$ particles (or replicas) are simulated according to a stopped Markov dynamics. The particle with minimal score, defined by the maximum of a level function over its trajectory is killed. Then a randomly chosen other particle is split in two and one offspring is re-sampled, starting from the entrance time of the current score. The method can estimate the probability of reaching a given score, as well as the distribution of the process conditioned by the latter event. In this paper, we prove a Central Limit Theorem on such estimators in the large sample size limit, using a recent similar CLT result we have obtained in~\cite{cdgr2} for Fleming-Viot type particle systems.

5: \medskip

6: 

7: \noindent {\em Index Terms} --- Sequential Monte Carlo, Fleming-Viot particle systems, Rare events simulation\medskip

8: 

9: \noindent {\em 2010 Mathematics Subject Classification}: 82C22, 65C05, 60K35, 60J60}

10: 

11: \end{abstract}

12: