Multi-objective Jaya Algorithm for Solving Constrained Multi-objective Optimization Problems

Abstract

Solving multi-objective optimization problems (MOPs) is a challenging task since they conflict with each other. In addition, incorporation of constraints to the MOPs, called CMOPs for a short, increases their complexity. Traditional multi-objective evolutionary algorithms (MOEAs) treat multiple objectives as a whole while solving them. By doing so, fitness assignment to each individual is difficult. In order to overcome this difficulty, in this paper, multiple populations are considered for multiple objectives to optimize simultaneously and a hybrid method of Jaya algorithm (JA) and quasi reflected opposition based learning (QROBL), to maintain diversity among populations, is used as an optimizer to solve CMOPs. An archive is also used to store all non-dominated solutions and to guide the search towards the Pareto front. A local search (LS) method is performed on the archive members to improve their quality and converge to the Pareto front. The whole above process is named as CMOJA. The obtained results are compared with the state-of-the-art algorithms and demonstrated that the proposed hybrid method has shown its superiority to its competitors. � 2020, Springer Nature Switzerland AG.