A Multi-Objective Bisexual Reproduction Genetic Algorithm for Computer Network Design

This research is concerned with a genetic algorithm focusing on the Pareto based approach for solving multiobjective optimization problems. Multi-Objective Bisexual Reproduction Genetic Algorithm (MOBRGA) is proposed herein. MOBRGA uses a concept of sexual selection with different types and mutation rates in reproducing offspring. The comparisons among MOBRGA and other algorithms were performed for evaluation purposes. The results showed that the proposed algorithm performed well in some benchmark multi-objective optimization problems. In Shaffer’s and Murata’s problems, MOBRGA could find the Pareto front which is the line of the nondominated solutions to the problems. The proposed algorithm was also tested with a computer network design, which is a multi-objective optimization problem. MOBRGA could successfully find network design solutions that meet user’s requirements.