Two-archive method for aggregation-based many-objective optimization

In this paper, a novel two-archive method is proposed for solving many-objective optimization problems. Our aim is to exploit the advantages of using two separate archives to balance the convergence and diversity. To this end, two updating strategies based on the aggregation-based framework are presented and incorporated into the two-archive method. In addition, we further extend this method by eliminating the restricted neighbourhood models. The proposed algorithms have been tested extensively on a number of well-known benchmark problems with 3-20 objectives. Experimental results reveal that the proposed algorithms work well on the many-objective optimization problems with different characteristics.