Decomposition Based Multi-objective Genetic Algorithm (DMOGA) with Opposition Based Learning

Multi-objective evolutionary algorithm has two goals i.e. diversity and convergence while solving MOP (Multi Objective Problem). These two goals can be achieved by proper selection of solutions. Real difficulty is selection of solution in presence of multiple conflicting objectives. MOP can be solved either by considering MOP as a whole or by using decomposition methods which solves scalar optimization sub problems simultaneously by evolving a population of solutions. This paper proposes decomposition based multi-objective genetic algorithm with Opposition operation. In this work Opposition Based Learning (OBL) concept is used in a unique way for weight vector generation. Also to have diversity among solutions and proper exploration of search space opposition based learning concept is used for population initialization and both parent and opposite parent are allowed to reproduce. The performance of the proposed methods is investigated on problems of CEC09 test suit. The experiments conducted show that OBL improves the performance of decomposition based Multi-objective Genetic Algorithm (DMOGA).