Related Experiment Videos
MM3 parametrization of four- and five-coordinated rhenium complexes by a genetic algorithm--which factors influence
Thomas Strassner1, Markus Busold, Wolfgang A Herrmann
1Anorganisch-chemisches Institut Technische Universität München, Garching b. München, Germany. thomas.strassner@ch.tum.de
Journal of Computational Chemistry
|April 2, 2002
Summary
Genetic algorithms (GA) optimize computational chemistry force field parameters. The Simple GA with a Tournament Selector provided the best results, enabling accurate prediction of rhenium compound structures.
Area of Science:
- Computational Chemistry
- Cheminformatics
- Bioinformatics
Background:
- Force field parameter optimization is crucial for accurate molecular modeling.
- Genetic algorithms (GA) offer a powerful approach for complex optimization problems.
- Previous methods often lack efficiency and accuracy in parameterization.
Purpose of the Study:
- To investigate the efficacy of genetic algorithms (GA) for optimizing molecular mechanics force field parameters.
- To determine the optimal GA configuration for computational chemistry applications.
- To develop new MM3 parameters for rhenium compounds.
Main Methods:
- Systematic evaluation of different GA compositions (Simple GA/Steady State GA, Roulette Wheel/Tournament Selector).
- Testing various crossover (0.5-0.9) and mutation rates (0.01-0.20).
- Development of new MM3 parameters for rhenium compounds using CSD structures.
Main Results:
- The Simple GA with a Tournament Selector demonstrated superior performance.
- New MM3 parameters were successfully derived for rhenium compounds (coordination numbers 4 and 5, oxidation states +V to +VII).
- GA-optimized parameters accurately reproduced experimental X-ray geometrical data, outperforming UFF calculations.
Conclusions:
- Genetic algorithms are highly effective for optimizing force field parameters in computational chemistry.
- The developed MM3 parameters enhance the modeling of rhenium-containing compounds.
- GA-based parameterization offers a significant advantage over traditional methods like UFF.