A Study on Many-Objective Optimization Using the Kriging-Surrogate-Based Evolutionary Algorithm Maximizing Expected Hypervolume Improvement

Table 2

The test problems and key properties.

Problems			Properties
DTLZ1	3, 5, 8, 10, 12, 15	+ 4	Multimodal, separable
DTLZ2	3, 5, 8, 10, 12, 15	+ 9	Concave, separable
DTLZ3	3, 5, 8, 10, 12, 15	+ 9	Multimodal, concave, and separable
DTLZ4	3, 5, 8, 10, 12, 15	+ 9	Concave, separable
DTLZ5	3, 5, 8, 10, 12, 15	+ 9	Concave, separable
DTLZ6	3, 5, 8, 10, 12, 15	+ 9	Concave, separable
DTLZ7	3, 5, 8, 10, 12, 15	+ 19	Discontinuous, separable
DTLZ7-10	3, 5, 8, 10, 12, 15	+ 9	Discontinuous, separable

is the number of objectives, and is the number of variables.