Multiphase Simulated Annealing Based on Boltzmann and Bose-Einstein Distribution Applied to Protein Folding Problem

Area of Science:

• Computational Biology
• Biophysics
• Algorithm Development

Background:

• The Protein Folding Problem (PFP) is a complex challenge in computational biology.
• Traditional Simulated Annealing (SA) algorithms face limitations in efficiently solving PFP instances.
• Developing advanced algorithms is crucial for accurate protein structure prediction.

Purpose of the Study:

• To propose a new hybrid Multiphase Simulated Annealing Algorithm using Boltzmann and Bose-Einstein distributions (MPSABBE).
• To enhance the efficiency and accuracy of solving Protein Folding Problem instances.
• To investigate the synergistic benefits of combining different statistical distributions in optimization algorithms.

Main Methods:

• Introduction of a four-phase algorithm: Multiquenching Phase (MQP), Boltzmann Annealing Phase (BAP), Bose-Einstein Annealing Phase (BEAP), and Dynamical Equilibrium Phase (DEP).
• Utilizing Boltzmann and Bose-Einstein distributions within simulated annealing search procedures.
• Employing a least squares method in the DEP for stochastic equilibrium detection.
• Parameter tuning using an analytical method considering maximal and minimal problem instance deterioration.

Main Results:

• MPSABBE demonstrated improved performance on various Protein Folding Problem instances compared to classical SA.
• The combined use of Boltzmann and Bose-Einstein distributions yielded superior results over using only the Boltzmann distribution.
• The algorithm's phases effectively manage temperature transitions and solution acceptance criteria.

Conclusions:

• The proposed MPSABBE algorithm offers a more effective approach for tackling the Protein Folding Problem.
• Hybridizing Boltzmann and Bose-Einstein distributions in SA significantly enhances optimization capabilities.
• This study highlights the potential of advanced statistical distributions in computational biophysics and protein structure prediction.