The Impact of Local Search on Protein-Ligand Docking Optimization

Common evolutionary approaches to protein-ligand docking optimization use mutation operators based on Gaussian and Cauchy distributions, with local search hybrids. The choice of a local search method is important for an efficient algorithm. We investigate the impact of local search with mutation operators by performing a locality analysis. High locality means that small variations in the genotype imply small variations in the phenotype. Results show that local search hybrids reduce locality and act as local optimizers with the solution as a starting point.