pestpp-glm unable to calculate super parameters when restarting with base_jacobian() #355
Description
Hello,
I'm using pestpp-glm.
In my control file I specified max_n_super, n_iter_base(-1), n_iter_super and had a run going. The run successfully carried out the base jacobian run and then shifted to the super parameter runs. After the second iteration of the super parameters it shifted to the central derivate and a lot of the base parameter were hitting their bounds. I stopped the run, modified the base parameter bounds (expanded the bounds) and then kicked off the run again, using the jacobian file created by the first iteration (w/ base_jacobian()) and specified the hotstart_resfile.
The run picked up and carried out 10 lambda upgrade runs, however when it tried to start the super parameter runs I got a bunch of warnings (example below):
... warning: failed to compute parameter derivative for SUP_195
warning: failed to compute parameter derivative for SUP_196
warning: failed to compute parameter derivative for SUP_197
warning: failed to compute parameter derivative for SUP_198...
.... can not compute super-par jacobian without the following parameters going out of bounds..
SDVA::iteration_jac() WARNING: parameter par1 already in frozen par names
SDVA::iteration_jac() WARNING: parameter par2 already in frozen par names
SDVA::iteration_jac() WARNING: parameter par3 already in frozen par names.....
Terminating super parameter iterations.
Max number of iterations to freeze parameters to compute jacobian exceeded....
The run then reverted back to the base parameter solution.
Is there an issue with the workflow that would cause this issue? If not, do you have any ideas why pest wasn't able to set up the super parameters?
Thanks!