A Differential Evolution Approach for Protein Folding Using a Lattice Model

Protein folding is arelevant computational problem in Bioinformatics, for which manyheuristic algorithms have been proposed. This work presents amethodology for the application of differential evolution (DE) tothe problem of protein folding, using the bi-dimensionalhydrophobic-polar model. DE is a relatively recent evolutionaryalgorithm, and has been used successfully in several engineeringoptimization problems, usually with continuous variables. Weintroduce the concept of genotype-phenotype mapping in DE in orderto provide a mapping between the real-valued vector and an actualfolding. The methodology is detailed and several experiments withbenchmarks are done. We compared the results with other similarimplementations. The proposed DE has shown to be competitive,statistically consistent and very promising.