I (we?) always assumed that using the same stopping criteria for each method was the "fair" way to compare optimisers, but e.g. Gary sees them as part of the method, so that the correct way to compare is to use the ones provided by the method.
So best to add a switch that allows both
I (we?) always assumed that using the same stopping criteria for each method was the "fair" way to compare optimisers, but e.g. Gary sees them as part of the method, so that the correct way to compare is to use the ones provided by the method.
So best to add a switch that allows both