NBI and MOGA-II, two complementary algorithms for Multi-Objective optimizations

Abstract

The NBI-NLPQLP optimization method is tested on several multi-objective optimization problems. Its performance is compared to that of MOGA-II: since NBI-NLPQLP is based on the classical gradient-based NLPQLP, it is fast and accurate, but not as robust, in comparison with the genetic algorithm. Furthermore a discontinuous Pareto frontier can give rise to problems in the NBI´s convergence. In order to overcome this problem, a hybridization technique coupled with a partitioning method is proposed.