A comparison between steepest descent and non-linear conjugate gradient algorithms for binding energy minimization of organic molecules

Abstract

The main intention of optimization is to bring about the “best” of any model by prioritizing the needs along with a given set of constraints. There is a wide range of optimization problems, among which, unfortunately, the problems that are formulated from nature are not convex in nature. Solving non-convex optimization problems are quite trickier than the conventional method of derivatives. One such problem is Computing the minimum value of binding free energy of various molecules. Minimization of free energy of a molecule is highly significant in the field of Molecular mechanics which is the foundation of computational biology. For any molecule, free energy refers to the amount of energy needed to separate an individual particle from a system of particles or to disseminate all particles in a system. The significance of free energy is that it can be used to compute the lowest energy conformation, which corresponds to the least Steric energy. Hence, this paper aims at computing the minimum value of binding free energy of various organic molecules in isolated conditions using the steepest descent algorithm and conjugate gradient algorithm and comparing them.