Closed fonsp closed 4 years ago
In my small experiments: before this fix, and before realising that domain
and dt
should be set to match each other, I was getting order-2 differences in the total control cost. Afterwards, I got very close results for different dt
s between 1
and 20
!
Good to know, thanks!
Thanks for catching this bug! I'll rerun the examples that used a non-standard dt (e.g. Replicate Belaia) and see if it affects the results at all.