Effects of Mutations on the Performance of Genetic Algorithms Suitable for Protein Folding Simulations

An implementation of a Genetic Algorithm suitable for protein folding simulations is analyzed. We focus on the role of mutations in the algorithm. In a simple folding model, it is shown that mutations are more important than in more conventional applications of genetic algorithms. To achieve good performance, mutations have to be introduced at a high rate under strict acceptance control.