Comparison of Various Mutation Operators of the Bacterial Memetic Algorithm on the Traveling Salesman Problem

Ákos Holló-Szabó, János Botzheim

Abstract: Bacterial memetic algorithms (BMA) are a relatively new branch of evolutionary algorithms that are good alternatives to traditional genetic algorithms. They meant a significant breakthrough regarding order-based optimizations like the traveling salesman problem. In BMA the selection and mutation steps are replaced by bacterial mutation. During bacterial mutation each specimen of the population is improved separately by cloning them, mutating each clone, and then selecting one as a replacement for the original specimen. The bacterial mutation is one of the most innovative and critical steps of the BMA. Bacterial mutations have many variants, but documentation lacks detailed statistics and behavior models. In this publication, we compare different bacterial mutation operators. The gathered data and observations will support all future research targeting BMA.