Posted by Evgeniy on February 22, 2009 at 17:02:10:
Dear Developers and Users of Molcas,
I am attempting to optimize geometry using CASSCF + numerical gradients but it does not converge. I use RFO method and the CARTESIAN option in alaska. I wonder whether it is worth trying some other optimization method or the problem lies elsewhere, e.g. in the use of numerical gradients or in the large size of the system (there are 28 atoms). I will appreciate any points regarding this problem. Thanks!