Fletcher-Reeves based Particle Swarm Optimization for prediction of molecular structure

The determination of the most stable conformers of a molecule can be formulated as a global optimization problem. Knowing the stable conformers of a molecule is important because it allows us to understand its properties and behavior based on its structure. The most stable conformation is that involving the global minimum of potential energy. The problem of finding this global minimum is highly complex, and is computationally difficult because of the number of local minima, which grows exponentially with molecular size. In this paper, we propose a hybrid approach combining Particle Swarm Optimization (PSO) and the Fletcher-Reeves algorithm to minimize the potential energy function. The proposed hybrid algorithm is applied to a simplified molecular potential energy function in problems with up to 100 degrees of freedom and also to a realistic potential energy function modeling a pseudoethane molecule. The computational results for both the cases show that the proposed method performs significantly better than the other algorithms.