J C Nash wrote:
> Extreme scaling quite often ruins optimization calculations. If you think
> available methods are capable of doing this, there's a bridge I can sell you
> in NYC.
Given that optim()seems to have problem with denormals but work correctly with
values greater than 1e-308, I thin
FWIW, I suspect the problem might have something to do with:
> 1/1e-308
[1] 1e+308
> 1/1e-309
[1] Inf
> 1e-309 > 0
[1] TRUE
> 1e-324 > 0
[1] FALSE
That is 1e-309 starts being treated as zero by division, but not by comparison
(which happens at 1e-324). This is not quite the range of values you
Thanks for all of your replies.
I would like to clarify that I am not expecting optim to return a good
result in these cases, as I am aware of the difficulties that come with the
numerical precision for such small numbers. I would be happy for optim to
take zero steps and return the initial value
The optim help page mentions scaling in the discussion of the "control"
argument. Specifically under the parscale description:
"Optimization is performed on par/parscale and these should be
comparable in the sense that a unit change in any element produces about
a unit change in the scaled va
Extreme scaling quite often ruins optimization calculations. If you think
available methods
are capable of doing this, there's a bridge I can sell you in NYC.
I've been trying for some years to develop a good check on scaling so I can
tell users
who provide functions like this to send (lots of)
Collin,
It is interesting that you get such strange results when passing denormal
values to optim(). Several possibilities for this spring to mind.
A better question though might be, "What kind of results are you expecting
when you pass such small values to optim()?" In general, you shouldn't
h
Às 17:30 de 23/12/2022, Collin Erickson escreveu:
Hello,
I've come across what seems to be a bug in optim that has become a nuisance
for me.
To recreate the bug, run:
optim(c(0,0), function(x) {x[1]*1e-317}, lower=c(-1,-1), upper=c(1,1),
method='L-BFGS-B')
The error message says:
Error in op
Hello,
I've come across what seems to be a bug in optim that has become a nuisance
for me.
To recreate the bug, run:
optim(c(0,0), function(x) {x[1]*1e-317}, lower=c(-1,-1), upper=c(1,1),
method='L-BFGS-B')
The error message says:
Error in optim(c(0, 0), function(x) { :
non-finite value supp
