EPCs with structured loops can be simulated in 7 but not in 9.6

The situation is worse than I expected. It seems there are more cases showing backward incompatibility of 9.6. Trying the simulate other models from ARIS demo database  7.2.5 (EPCs with ERM routing but no feedback loops) gives again "unexpected exception".

What version are you using exactly? Do you have a sample model you could share?

I cannot reproduce the exception with 96.0.3.914155 and a very simple EPC which contains only the loop structure.

Regarding the semantic check, that's really only for ARIS 6 and before, so you can ignore its results.

Ralf,

I'm using 96.0.3.906056 with JRE 1.7.0_67

After restoring the "United Motors Group" database of 7.2.5, simulating for example model UMC.21247 gives "unexpected exception". The same model and all the rest (UMC.21243, UMC.21245) in the group run without problem in 7.2.5. With the rest of the simulation demo models the situation is:

• 87384727-2897-11e0-2b1c-0025b3bf9d9d - OK
• 4a31a090-e14d-11dc-1cf0-00164157f4fd - OK and sometimes: "Error code 4 - Could not find alive instance with specified type or id"
• 3a7336e9-1a57-11d3-ad40-00a024e7ef08 - OK and sometimes: "Error code 4 - Could not find alive instance with specified type or id"
• 026d75c1-5bae-11e0-21b5-5c260a227a4f -  OK and sometimes: "Error code 4 - Could not find alive instance with specified type or id"

It might be that the Error code 4 is related with different state of the simulation server. I've no idea.

For the model with the loop, I can make a shareable version. Would it be OK if I upload it here as XML?

By the way, Erlang distribution doesn't seem to work in both 7.2.5 and 9.6. That's not so important but I would be glad to know, is it a known bug or the problem is at my end.

Okay, I could reproduce the exception with the model in the UMC database and reported it as a bug. Obviously this is related to the ERM attributes.

Uploading the sample model as XML should be okay. Does it contain ERM attributes as well? Because if it does, then the exception likely has the same cause as in case of the model from the UMC database.

And the Erlang distribution not working is a long standing bug. I'll try to revive the bug report...

Thanks for reporting all of this.

Yes, indeed it is with ERM attributes as well. I have just removed them I tested the same flow structure but with probability routing and it works. However, in this particular case and most cases for simulating iterations, the conditions needs to be specified with ERM attributes.

(I have more issues with simulation; I wonder is it better to continue this thread or submit separate entries?)

If it's better to start separate threads probably depends on the nature of the issues. If they concern the usage of simulation, how to create models for simulation, and so on, then separate entries could make sense because such topics usually require a bit of discussion. If they concern bugs in the simulation, then we can continue here. Note however, that bugs are ideally reported to ARIS Support.