Closed leonlan closed 2 years ago
I'm closing this PR because the (new) benchmark results surprisingly show that there is not much to gain, even though earlier experiments showed slight benefits (-30 cost). I don't think it's worth my time to continue experimenting with this.
Closes #96.
I'll be performing some runs on the cluster to find the best mutation parameters.