Model debugging - repeat dashboard errors & treenode corruption

Question

question

Christopher Morgan asked Nov 7, '18 Cliff King answered Nov 9, '18

Model debugging - repeat dashboard errors & treenode corruption

After adding in a few dashboards and using the explore function under experiment, i'm getting a lot of exceptions when running the model with the experiment control explore function under the various scenarios i've developed for testing. This happens anytime i switch between one of the dashboard tabs while the model runs.

I've also periodically been having tree node errors for different nonselectable objects (floors) in my model where the variables->objectref will consistently become corrupted on reset to a non english character of a. I've been experiencing this a lot, particularly when resetting the model while its running

What could be causing these errors and what can i do for the future to prevent/debug them?

dashfix.fsm

Software Version:

FlexSim HC 5.3.6

hc debugging

dashfix.fsm (6.1 MiB)

Answer 1 · 2018-11-09T20:00:06Z

Cliff King answered Nov 9, '18

Sorry for the frustration you're having with dashboards. The sheer number of calculations that take place when dashboards need to be updated is astronomical and can make them temperamental to say the least. Sometimes the problems can be resolved by deleting and recreating the dashboards and/or widgets, but in general you need to be careful in making any edits while the model is running or with a simulation time greater than 0. Here's some general tips when working with dashboards that will help.

Keep all but one dashboard window closed when doing manual runs and close all dashboards when running an experiment.
Line graphs create a horrendous amount of data (and post calculations), so use sparingly.
When you want to run a specific experiment scenario manually (ie. with the Explore tab of the Experimenter window), always reselet the Scenario and maybe even the Replication from the dropdown list box even if the scenario and replication number you want to run is showing in the field. The act of choosing a scenario runs the script associated with Experiment Variables defined for the given scenario on the Scenarios tab of the Experiment window.
Use a different set of dashboard widgets for displaying the results of the last single model run than you use to compare scenarios as defined by the Experimenter. Your Sample Sets will be different.
During an experiment run, do not touch your keyboard or mouse until the experiment has completely finished running and you see the popup window indicating a successful experiment run. Then open one dashboard at a time (remember you played it safe and closed all your dashboards before running your experiment) and WAIT after you've opened a dashboard to let the widgets on the dashboard completely finish refreshing (don't touch anything while you wait).
Sometimes you can clear out window and/or dashboard related errors by opening the default workspace using the main menu: View > Open Default Workspace.
If you ever see the phrase "exception: Exception Caught in updatestates()" in the System Error Console window, the state of your current model has been corrupted and there's nothing you can do to save the model (this is why you should always save your model before running it). It's best to close FlexSim if this happens and start things up again.

Hopefully these tips will save you some frustration in the future. Let us know if you have further questions.

question