Multi-objective Constrained Genetic Algorithm Based on Pareto and Hierarchical Sorting

In order to further improve the computational efficiency, NSGA-II algorithm was improved from three aspects of non-dominated set construction, individual ordering and new population generation in this paper. Firstly, the population was divided into feasible population and infeasible population, feasible and infeasible population individual respectively using non-dominated sorting and mixed sorting to construct sorting set, and a new population generation method was established which only calculates the crowding distance of individuals for a specific sorting set. Furthermore, the framework of multi-objective constrained genetic algorithm based on Pareto and hierarchical sorting was constructed, which could reduce the calculation time of non-dominated set and individual crowding distance of NSGA-II. Finally, the algorithm was verified by three examples, and a satisfactory Pareto front was obtained.