A Improved NSGA-II Algorithm for Constrained Multi-objective Optimization Problems


Constrained multi-objective optimization problems (CMOP) have been a research focus in multi-objective optimization problems (MOP). Based on the technologies from NSGA-II such as non-dominated sorting, elitist strategy and niche technique, this paper proposes an improved NSGA-II algorithm for constrained multi-objective optimization problems. In the improved algorithm, a partial order relation is first set up. Then according to the partial order relation, the individuals are sorted for generating the non-dominated individuals. Moreover, to enhance the evolution's ability, some individuals are evolved in the same generation and the two-body crossover is adopted. In addition, non-dominated individuals generated in each generation are archived to Pareto set filter to reserve all individuals with good characteristic generated in the evolving process. Finally, some Benchmark functions are used to test the algorithm performance. Test result shows the availability and the efficiency of the algorithm.