Enhance Version of Genetically Adaptive Multi-algorithm for Multi-objective Optimization Problems

Multi-objective evolutionary algorithm (MOEAs) are well established population based stochastic techniques. They employ various evolutionary operators and approximate a set of optimal solutions for the problem at hand in single run unlike traditional mathematical programing. Each search operator have certain benefits and limitations. A multiple use of operators with self-adaptive procedure can improve the performance of particular MOEAs. This paper suggests the enhance version of a genetically adaptive multi-algorithm for Multi-objective (AMALGAM) and examined of their performance upon two different types of test suites of continuous multi-objective optimization problems, the widely used ZDT test problems and the test instances which are recently designed for the special session of MOEAs competition in Congress of Evolutionary Computing 2009 (CEC’09). The suggested algorithm have discovered better approximated set of optimal solution for almost all benchmark functions. We have used inverted generational distance (IGD) as a performance indicator and the proposed algorithm have obtained minimum average IGD-metric values as compared to the NSGA-II that is commonly used in comparative analysis.