PSO Method for Fitting Analytic Potential Energy Functions. Application to I^-(H₂O).

Area of Science:

• Computational Chemistry
• Theoretical Chemistry
• Molecular Modeling

Background:

• Accurate potential energy functions are crucial for molecular simulations.
• Fitting these functions to ab initio calculations can be computationally intensive.
• The interaction between iodide (I-) and water (H2O) is important in various chemical systems.

Purpose of the Study:

• To develop and validate an efficient method for fitting analytic potential energy functions to intermolecular potential energy curves.
• To compare the performance of Particle Swarm Optimization (PSO) with a genetic/nonlinear least-squares algorithm for this fitting task.
• To assess the accuracy and computational cost of different molecular models for I-(H2O) interactions.

Main Methods:

• Utilized Particle Swarm Optimization (PSO) to fit a six-parameter analytic potential energy function to DFT/B97-1 calculated potential energy curves for I-(H2O).
• Employed two models for H2O: a three-site model and a four-site model including a ghost atom.
• Compared PSO fitting results with those obtained using a genetic/nonlinear least-squares algorithm.

Main Results:

• The four-site model significantly improved fitting accuracy compared to the three-site model for both algorithms.
• PSO demonstrated superior accuracy and reduced computation time compared to the genetic/nonlinear least-squares algorithm.
• PSO achieved root-mean-square errors (RMSE) of 1.37 and 0.22 kcal/mol for the three- and four-site models, respectively, with manageable computation times.

Conclusions:

• Particle Swarm Optimization is a highly efficient and accurate method for fitting analytic intermolecular potential energy functions.
• The four-site model with PSO provides an adequate representation of I-(H2O) interactions for calculating cluster properties.
• PSO offers a robust and time-saving alternative for developing accurate molecular potentials.