1: \begin{abstract}

2: We apply a recently developed adaptive algorithm that systematically improves the efficiency of 

3: parallel tempering or replica exchange methods in the numerical simulation of small proteins. 

4: Feedback iterations allow us to identify an optimal set of temperatures/replicas which are

5: found to concentrate at the bottlenecks of the simulations. 

6: A measure of convergence for the equilibration of the parallel tempering algorithm is discussed. 

7: We test our algorithm by simulating the 36-residue villin headpiece sub-domain HP-36 where 

8: we find a lowest-energy configuration with a root-mean-square-deviation of less than $4$ \AA\ 

9: to the experimentally determined structure. \\ \newline

10: 

11: %\noindent

12: %{\bf Corresponding author}\\

13: %Ulrich H.E. Hansmann \\

14: %John-von-Neumann Institute for Computing \\

15: %Forschungszentrum J\"ulich \\ 

16: %D-52425 J\"ulich, Germany \\

17: %Phone:	+49-2461-61-1526 \\

18: %Fax:		+49-2461-61-2430 \\ 

19: %E-mail:	u.hansmann@fz-juelich.de \\ \newline \newlinex

20: %{\bf Manuscript information}\\

21: %text pages: 18\\

22: %words in abstract: 90\\

23: %characters in paper: approx. 40'000\\

24: \end{abstract}

25: