EXHAUSTIVE CONFORMATIONAL SEARCH AND SIMULATED ANNEALING FOR MODELS OF LATTICE PEPTIDES

We consider simple lattice models for short peptide chains whose states can be exhaustively enumerated to find the lowest energy conformation. Using these exact results and numerical simulations, we compute the distributions for the mean time t(N), required to find the global minimum energy state by simulated annealing (SA), as a function of N, the number of units in the chain. On the basis of scaling arguments, the time t(N), to find the global minimum energy of longer chains, beyond the range covered by exhaustive enumeration, can be estimated. On the basis of the observed exponential increase in folding time of the standard SA algorithms, it is imperative that better algorithms be found for minimizing longer chains.