Left-Right crowding distance (LRCD) calculation method in NSGA2 to preserve diversity distribution

Recent years, Evolutionary Algorithms (EAs) have become the most practical method in solving Multi-Objective Problems (MOPs). The diversity performance is one of the most important parts in assessing how EAs perform in Multi- Objective problems. In this paper, we propose a new crowding distance calculation method (we call it LRCD) relative to the original NSGA2. Six Multi-Objective problems have been tested to illustrate the performance of the new version of NSGA2, we find the new method performs better in some problems in diversity preservation but there is no clear improvement in other problems. Furthermore, tests have been done to illustrate the new method´s robustness.