Preference-inspired Co-evolutionary Algorithms

Pareto approximation data

These are the Pareto approximation sets obtained by PICEA, HypE, MOEA/D, eMOEA, NSGA-II and rand for WFG test problems(WFG2 to WFG9) with 2, 4, 7 and 10 objectives, respectively. All the algorithms are performed 31 independent runs within 25000 function evaluation times.

Downloadable zip file (37MB): Pareto approximation data

Hypervolume results

The hypervolume results computed based on the offline archive (filtered by a cluster method applied in SPEA (Zizlter, 1999)) for WFG test problems(WFG2 to WFG9) with 2, 4, 7 and 10 objectives, respectively. The results are based on 50 independent runs within 25000 function evaluation times.

Downloadable zip file (2.6MB): Hypervolume results

Attainment surfaces

Attainment surfaces (colour reproduction) for 2-objective WFG test instances (WFG2-2 toWFG9-2)

Downloadable zip file (1.1MB): Attainment surfaces (colour reproduction)