VMD-L Mailing List
From: Mayne, Christopher G (cmayne2_at_illinois.edu)
Date: Mon Mar 07 2016 - 12:24:58 CST
There is no absolute threshold below which the results are satisfactory. The objective function does not take into account the number of terms that are included in the optimization; therefore, the objective function is much more likely to be larger for optimizations that contain more terms than for those that contain very few terms. Typically, I run iterative downhill optimizations (changing my output LOG filename) until the objective value no longer decreases.
> On Mar 2, 2016, at 8:36 PM, Soumya Lipsa Rath <soumyalipsabt_at_gmail.com> wrote:
> Dear Vmd Users,
> During bond optimization, under the results section we get Current Final Obj. value and Previous Final Obj. value. What is the satisfactory Obj. Value (values) after which one can use the refined bond/angle parameters. Should it also converge like the COLP or should it be zero?. Although it has been enquired earlier in vmd mail list, I failed to find a definite answer. Kindly provide me some insights for the same.