Experimental Analysis of 2-multiobjective Evolutionary Algorithm

A multi-objective optimization problem (MOP) is often found in real-world optimization problem. Among various multiobjective optimization techniques, multi-objective evolutionary algorithm (MOEA) is highlighted as a good candidate due to its flexibility, feasibility, and its ability to handle multiple solutions. Among various MOEAs, we analyze 2MOEA which can achieve good convergence and diversity using 2-dominance. We compare 2-MOEA with controlled elitist non-dominated sorting genetic algorithm (NSGA-II) which is currently one of the most popular and widely used MOEAs and have showed reasonable performance in various applications. We compare two algorithms experimentally for popular and scalable test function DTLZ2 and its modified version. The experimental results show that 2MOEA is better than NSGA-II in convergence and is comparable to NSGA-II in diversity.