[R-sig-ME] optimizers for mixed models

[Ben Bolker]

Ross Boylan <ross at ...> writes:

> 
> On 3/14/2013 10:55 AM, Ross Boylan wrote:
> > Optimization was originally via a custom optimizer using rho.  The 
> > custom optimizer did not incorporate bounds, and blew up when it got 
> > rho outside of [-1, 1].  So we switched to atanh(rho) as the target of 
> > optimization.  However, for some simulated datasets that failed to 
> > converge, as atanh(rho) marched slowly off toward infinity.  We 
> > switched to optim with bounds to cut that process off.
> >
> > So perhaps we should go back to rho, but using optim or the other 
> > bounded optimizers you suggested.
> >
> > So the fact that atanh(rho) is unbounded is a feature from some 
> > perspectives, but a bug from others. 
> I forgot to mention that we actually have analytic second (and first) 
> derivatives.  Switching to optim from the packages internal optimizer 
> meant we're no longer using the analytic 2nd derivative.

  Which package?

 A couple more points:

 * I don't know offhand which optimizing tools in R (1) allow for
box constraints and (2) can take both a user-specified gradient
and a user-specified hessian (there are several that take user-specified
gradients); I think there may be some in optreplace/optimx, but I'm not
sure.

 * according to Dave Fournier, AD Model Builder actually can do GH
quadrature with multiple terms per random effect (the documentation
is wrong/out of date), so you might look into that ...  (as I recall
this is a single RE with multiple terms, not crossed RE -- right?)

  Ben