Finding the needle in the Haystack: algorithms for conformal optimization

Abstract

Conformational optimization of many-body systems in molecular physics and biophysics remains an unsolved problem. In recent years, various algorithms that use potential smoothing to simplify the problem of isolating the global energy minimum have been devised. Although Gaussian transforms have led to important results, the method of bad derivatives, which requires function evaluations only and is derivative free, appears to be promising. Quantum-mechanical annealing is an optimization paradigm that should lead to a variety of optimization algorithms based on alternative representations of the system, such as path integrals, which are well suited for condensed-phase simulation.