JaamSim 2015-02 released

[Harvey Harrison]

Changes in this version:

- Simulation keyword 'SimulationTimeScale' has been replaced with 'TickLength' which allows you to directly specify
the length of the internal simulation steps.  The default has also been changed to be 1 microsecond.

- A bug has been fixed that forced the user to press the stop button when their model fails a validation check at startup,
now you can just fix your inputs and hit play again without the extra button press to restart things.

- The default value that appeared for Unit's keyword 'ConversionFactorToSI' has been fixed in the EditBox, this was a cosmetic
issue and had no effect on model results

- the Simulation functionality provided through the 'PauseTime' keyword now takes effect in all cases, including when the model
is run in batch mode.

- improved icons have been provided for some types that did not provide a good preview image before in the Model Builder

- Lots of refactoring in the Process Flow objects, I'll let Harry expand on new functionality there

- For those people coding their own objects, many of the most common simulation/eventmanager methods have had their
javadocs added or improved.  Please make requests for any methods you find lacking and I will prioritize those first.

Cheers,

Harvey

[Jerry L]

Thanks Harvey!  you're spoiling us. 

[Jerry L]

Should I worry about this new error message? (attached)  

[Jerry L]

My models are giving different results now...investigating.

[Harry King]

Jerry,

I was about to post a warning about this message. Yes, you can ignore it. Any keywords that JaamSim does not recognize are ignored.

Once you re-save your input file, the old "SimulationTimeScale" keyword will be gone and the input file will load correctly.

Harry

[Harvey Harrison]

No....unfortunately if you started your model from drag and drop, this would be added to the saved file, just remove that line from your inputs.

Harvey

On 16 Jan 2015 18:26, "Jerry L" <jerryle...@outlook.com> wrote:

Should I worry about this new error message? (attached)  

--
You received this message because you are subscribed to the Google Groups "Jaamsim Users Discussion Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to jaamsim-user...@googlegroups.com.
Visit this group at http://groups.google.com/group/jaamsim-users.
To view this discussion on the web, visit https://groups.google.com/d/msgid/jaamsim-users/6156a193-7378-41ce-bbc2-1ffed17ad3b2%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Harry King]

Jerry,

The difference is probably the change to the default time scale. If you first built you model with an older version of JaamSim from about 6 months ago, the default time scale was different. Check the value for the SimulationTimeScale input for Simulation. You can restore this value by setting the new keyword TickLength to (3600 s)/SimulationTimeScale. However, it is better to use the default precision of 1 millisecond to avoid round-off error in your model's time keeping.

Harry

[Jerry L]

I won't worry about SimulationTimeScale.

But I do have at least two models giving radically wrong results now.  One uses assembly and one just has regular queuing.. I am investigating now...

On Friday, January 16, 2015 at 6:35:15 PM UTC-5, Harvey Harrison wrote:

[Harry King]

Jerry,

The Assembly keyword Choice has been deleted -- you need to send your entities directly to the appropriate queue. I'll explain why this change was made later this evening.

The other thing that changed was the obj output for Server, etc. It is now the entity that is being served, not that last entity to arrive at the Server (and sent to the Queue).

Harry

[Jerry L]

I just sent Harry the models...  

Your post could explain the results.  Let me see....

[Jerry L]

So if I used Choice previously, I should now have a branch that sorts them to the queues?

[Jerry L]

For the Assembly model, putting a Branch before the Assembly worked.

Fortunately, it should only take 5 minutes each in 4 places.

On Friday, January 16, 2015 at 9:56:09 PM UTC-5, Jerry L wrote:

[Jerry L]

Yeah!  The one with the set-up time recorded on the first part, where the machine was giving the wrong time based on the last part in the queue.... That one is works now for the first time ever!! 

But another model that was basing thresholds on queue and conveyor Work-In-Process stopped working.  So something changed in how it reports queues... investigating....

So:

3 models stopped working (investigating)

2 stopped but now re-written to work

1 that never worked but now does

7 not impacted

 

[Jerry L]

[EntityConveyor].NumberAdded no longer seems to be counting.  This is straight from the generator to the conveyor 9no queues).

On Friday, January 16, 2015 at 6:35:15 PM UTC-5, Harvey Harrison wrote:

[Harry King]

Jerry,

We will get NumberAdded fixed for EntityConveyor early next week.  Sorry about that.

Harry