Closed erex closed 9 months ago
Re-run this simulation and got the following errors (note have not reproduced original error in 1500 reps):
Summary of warnings and errors:
simpleWarning in optimx.check(par, optcfg$ufn, optcfg$ugr, optcfg$uhess, lower, : Parameters or bounds appear to have different scalings.
This can cause poor performance in optimization.
It is important for derivative free methods like BOBYQA, UOBYQA, NEWUOA.
(Model number: 2) (occurred 2 time(s) in repetition(s): 76, 401)
simpleWarning in ddf.ds(model = dsmodel, data, meta.data = meta.data, control = control, : First partial hessian is singular; using second-partial hessian
(Model number: 2) (occurred 1 time(s) in repetition(s): 127)
-----
Next run produced:
> simulacion.paralela.3km <- run.simulation(simulacion.paralela.3km)
Summary of warnings and errors:
simpleWarning in optimx.check(par, optcfg$ufn, optcfg$ugr, optcfg$uhess, lower, : Parameters or bounds appear to have different scalings.
This can cause poor performance in optimization.
It is important for derivative free methods like BOBYQA, UOBYQA, NEWUOA.
(Model number: 2) (occurred 2 time(s) in repetition(s): 655, 686)
simpleWarning in ddf.ds(model = dsmodel, data, meta.data = meta.data, control = control, : First partial hessian is singular; using second-partial hessian
(Model number: 2) (occurred 1 time(s) in repetition(s): 767)
-----
I will move this to the list of things for the next release rather than close completely but for now I cannot replicate it.
Still cannot replicate so closing
Reported by a user on 20Sep22 that
dsims
fails with following messageI presume it is
mrds
, called byds
that callsoptimx.check()
.I'm not sure if
DS.analysis.R
handles this situation properly, because I'm not sure the event would be classified as either an error or a warning.Might be indicative of additional problems arising in
mrds
because optimx.check gets to this part of the code only if the parameter estimates have landed outside the bounds for that parameter. Not sure how that comes to pass.