Hybrid Evolutionary Algorithm for Molecular Geometric Optimization

Abstract

In this work a hybrid algorithm is developed to solve a geometric optimization problem whichis classified as NP-complete problem. The proposal effectively combines an evolutionary algorithm with a clustering algorithm to balance the exploration and exploitation of the search space. This algorithm works with the secondary structure of the molecule using the backbone dihedral angles j (phi) and y (psi) as the main components because de energy depend directly of them, the angles j and y are described in a Ramachandran map and the local search is guided towards the conformations of the lowest energy.