Average best colouring and standard deviation of each hyper heuristic

Fig. 6. The mean percentage utilisation rate of low level heuristics produced by the RL-ILTA and DH-ILTA hyper-heuristics on the COLOR02 benchmark instances.Figure optionsDownload full-size imageDownload as PowerPoint slide
4.4. Performance comparison to previously proposed approaches
Table 7.
Performance comparison for different graph colouring approaches based on the best result. The entries in bold indicates the best result obtained by the associated algorithm for WWL 70 given instance.Instancek∗k∗RL-ILTAM-LLELIMXGPX-LIGPX-CBLLE-eLLE-bLLEKir-BKir-CDSJC125.155–––––––––DSJC125.51717181818181919181918DSJC125.94444444444444444444545DSJC250.188999999999DSJC250.52829–313131–––3030DSJC250.97272757575747474747777DSJC500.11213–141414–––1414DSJC500.5485255––––––––le450_25a2525–252525–––2525le450_25b2525–252525–––2525le450_25c2527292828282929292828le450_25d2527–282828–––––Wins12133311122Full-size tableTable optionsView in workspaceDownload as Pacinian corpuscles CSV