Parallel Evolutionary Multi-Quenching Annealing for Protein Folding Problem
Keywords:Multi-quenching Annealing, Parallel Genetic Algorithms, Protein Folding Problem, Simulated Annealing
The Protein Folding Problem (PFP) consists in determining the functional three-dimensional structure or Native Structure (NS) of a protein, which normally has the lowest Gibbs energy. In this paper, a new hybrid Parallel Evolutionary Multi-Quenching Annealing Algorithm (PEMQA) is proposed to obtain high-quality solutions for the target proteins. PEMQA generates an initial population of solutions using a Genetic Algorithm (GA). Furthermore, a Multi-Quenching Algorithm (MQA) is executed in an independent core using each of these Genetic Algorithm (GA) solutions. A master process determines which MQA delivers the best solution. PEMQA uses shared memory parallel programming and is implemented in SMMP (Simple Molecular Mechanics for Proteins). The incorporation of evolutionary processes in a PEMQA algorithm allows an improvement in MQA capacity of exploration. Results obtained with PEMQA outperform most of those achieved by the classic SA reported in current state of the art literature.