Abstract
Multi-Objective Evolutionary Algorithms (MOEAs) have been successful in solving mathematical and real-world multi-objective optimization problems by evolving a set of optimal solutions, which are known as Pareto-optimal solutions. However, there are certain limitations with those algorithms such as slow convergence, lack of effective terminating condition to name a few. To address such challenges, hybrid MOEAs are being designed and studied where the global exploration power of MOEAs are combined with local exploitation modules of various numerical optimization techniques. However, hybridization itself brings new challenges in its implementation. In this work, a hybrid MOEA is presented in which random mutations are performed on the initial population to start with a better and diverse set of solutions. Moreover, a local search module is coupled to execute periodically on a least crowded non-dominated solution at a certain interval of generations. The proposed algorithm is tested on a set of benchmark multi-objective optimization problems and compared with the NSGA-II. The convergence plots demonstrate the superiority of the proposed algorithm over NSGA-II.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call