Hello, everyone.
I wanted to chime in to this conversation to say that we have, in fact, discussed options to help with non-convergence in the past, and we are very interested and appreciate your suggestions. We will discuss some of the above suggestions with our development experts.
As Richard indicates, what we want to maintain is your ability to apply your engineering knowledge and judgement when making decisions on non-convergence issues, rather than having the software automatically alter models without any reporting coming back to you (i.e., your model changed somewhere, you have no idea now, but now it magically converges? Sounds a bit scary to me!)
In addition to the great suggestions made, some future possibilities might be some sort of a wizard or a prompt that comes up that gives you common options/tips that help fix convergence issues. You could then go try those options to see if it remedies the problem.
MOST OF ALL, I think it would be important to then have a report or some indication on your output reports where things were changed to fix convergence issues--even if it is a log file of some sort that you can go back and look at to see what was changed. Or, as Michael suggests, something visual on the model to show where changes were made.
Does this match with some of what you are hoping for with these suggestions?
Disclaimer--this is not currently in development or planned for our product yet, but your feedback is key in us understanding what users need and would appreciate most in the future.
Technically speaking, solving convergence issues can be pretty complex and unique to each model/job, so that is what makes this request pretty challenging to solve. However, we are listening and discussing, so keep the feedback coming.
Stephanie Corken
CAESAR II Product Owner
Hexagon PPM (CAS)
stephanie.corken@hexagon.com
Edited by StephDonovan (05/15/18 11:26 AM)