Closed schrum2 closed 8 years ago
Made output of child files option and turned off by default. Still need to disable scores for each generation when there is only one objective, since these scores are only used for hypervolume calculation.
Made tracking of combining crossover information optional, which makes archetype management easier as well.
parent fitness and other scores from each generation are still logged, but I think these are worth keeping after all. The rest is done.
Decreasing the amount of I/O that the system performs could speed things up a bit. Things like the child population scores don't really seem necessary ... all that is usually needed are the parent scores. Even the individual parent scores for each generation could probably be dropped when there is only one objective ... all that score data is really only needed to calculate hypervolume for multiobjective runs.