Multi-objective genetic algorithm based on cloning mechanism and its application

Proposed a new clone multi-objective genetic algorithm (for short: NCMGA), which introduced clone mechanism based on NSGA-II. It has two main characters: for one thing, the superior individuals are reserved by twice selection, and then these individuals are cloned. With this method, it greatly enhances the information transmission between the parents and offspring, and improves population convergence effectively. For another, a single-point compound crossover operator is designed, which makes the crossover coefficient of the individuals have a dynamic connection with their ranks. The individuals with lower ranks are reserved, while those with higher ranks are crossed, which contribute to improve population diversity. To know how competitive NCMGA is, we compare it against NSGA-II by calculating four well-known benchmarking problems. The comparison results show that NCMGA clearly outperforms NSGA-II in the aspects of convergence and diversity. Finally the application of the cantilever beam bi-objective optimization example shows that the proposed multi-objective optimization algorithm in this work has high application value in the engineering.