RANGE: A robust adaptive nature-inspired global explorer of potential energy surfaces

Area of Science:

• Computational Chemistry
• Materials Science
• Bioinformatics

Background:

• Accurate chemical structure representation and global minima identification are crucial for computational chemistry and materials science.
• Exascale computing advancements enable more sophisticated methods for exploring potential energy surfaces.
• Swarm intelligence algorithms, like Artificial Bee Colony (ABC), show promise in optimization tasks.

Purpose of the Study:

• To develop a hybrid metaheuristic framework integrating ABC and Genetic Algorithms (GA) for robust global optimization.
• To create a scalable, Python-based tool named RANGE (Robust Adaptive Nature-inspired Global Explorer) for computational chemistry applications.
• To assess the performance of RANGE against standalone ABC and GA algorithms.

Main Methods:

• Developed a hybrid metaheuristic framework combining ABC's exploration with GA's exploitation.
• Implemented the framework in a scalable, Python-based tool, RANGE, with interfaces for potential energy evaluators.
• Evaluated RANGE's performance on molecular clusters and heterogeneous surfaces.

Main Results:

• RANGE demonstrated superior efficiency and robustness compared to ABC- or GA-alone algorithms.
• The hybrid approach effectively addressed challenging global optimization problems.
• Successful application across diverse chemical systems, including molecular clusters and surfaces.

Conclusions:

• The RANGE tool offers a powerful and versatile solution for global optimization in computational chemistry and materials science.
• Hybrid metaheuristic approaches combining exploration and exploitation are highly effective.
• RANGE is well-suited for high-performance computing environments, facilitating complex chemical structure discovery.