[ADMB Users] Random effects failure in the inner optimization

[dave fournier]

I've been working with Anders  on better recovery when the Hessian
is not positive definite in random effects models. I have attached 
modified code
with a simple example which fails with the current code and succeeds with
these patched versions.  It certainly won't handle all difficult models, but
it is a start in this direction.



-------------- next part --------------
A non-text attachment was scrubbed...
Name: code.zip
Type: application/zip
Size: 12084 bytes
Desc: not available
URL: <http://lists.admb-project.org/pipermail/users/attachments/20110405/176dcbe9/attachment.zip>