A Two-Archive Algorithm with Interactive Mechanism for Many-Objective Optimization

Table 2

Performance comparison on DTLZ1-4 problems with respect to the average HV values.

Problem	Obj.	iTwoArch	TwoArchA+	TwoArch2	MOEAD	MOEADAM2M	NSGAIII
DTLZ1	3	0.973505	0.973558	0.975170	0.973651	0.970025	0.972512
	5	0.998961	0.988324	0.997340	0.996220	0.998330	0.998628
	8	0.999931	0.999914	0.999242	0.999560	0.999926	0.999604
	10	0.999999	0.999775	0.998868	0.999957	0.999988	0.999731
	15	0.999997	0.999991	0.974723	0.999227	0.988791	0.999710
DTLZ2	3	0.926717	0.926697	0.929175	0.926730	0.926001	0.926643
	5	0.990526	0.990386	0.983421	0.986239	0.990349	0.990434
	8	0.999321	0.999302	0.993856	0.998597	0.999246	0.999259
	10	0.999918	0.999901	0.992931	0.999772	0.999875	0.999876
	15	0.999996	0.999998	0.981044	0.999624	0.999506	0.999996
DTLZ3	3	0.978073	0.997819	0.972712	0.978151	0.976368	0.977728
	5	0.998658	0.960642	0.992559	0.997989	0.998381	0.990501
	8	0.999954	0.996681	0.817973	0.999784	0.999956	0.993984
	10	0.999668	0.969624	0.000000	0.999886	0.999994	0.946079
	15	0.999965	0.999999	0.691400	0.999417	0.997129	0.999696
DTLZ4	3	0.926715	0.909868	0.929180	0.918331	0.926035	0.926660
	5	0.990541	0.990522	0.985053	0.985951	0.990448	0.990503
	8	0.999370	0.999297	0.994405	0.999350	0.999420	0.999355
	10	0.999925	0.999913	0.993294	0.999916	0.999932	0.999919
	15	0.999999	0.999995	0.983069	0.999992	0.999997	0.999999

Note. “” means that the result is significantly outperformed by iTwoArch.