Benchmarking Results
Comparing 15 scheduling algorithms across 7 datasets. Colors indicate mean makespan ratio (lower is better; 1.0 = best on every instance).
Loading interactive heatmap...
Methodology
Each cell shows the mean makespan ratio of a scheduler on a dataset. The makespan ratio for an instance is computed as the scheduler's makespan divided by the best makespan achieved by any scheduler on that instance.
A value of 1.0 means the scheduler always achieves the best makespan. Higher values indicate worse relative performance. The color scale ranges from green (1.0) to red (5.0+).
Results are from the SAGA benchmarking pipeline. See the paper for full details.