Re: [4diac-dev] Detecting self-cycles in ECC

[Alois Zoitl <alois.zoitl@xxxxxx>]

Hi,

as Martin said I also find this a good starting point. Based on that the next problematic self transition would be one with only input data elements in an event
less transition condition. However I'm not sure if this information is already available in the model.

BR,
Alois



On Thu, 2020-06-04 at 09:24 +0000, Martin Melik-Merkumians wrote:
> Hi,
>
> empty transitions are illegal as far as I know, so that would be an incomplete transition anyhow.
>
> The loop check with a 1 transition is a good idea, as this would result in a deadlock situation, as the FB would never stop executing blocking all other FBs on the resource and possibly flooding the event chain.
>
> Best regards,
> Martin
>
> --------------------------------------------------
> Dipl.-Ing. Martin Melik-Merkumians
> Advanced Mechatronic Systems
> Automation and Control Institute (ACIN)
> TU Wien
> DVR-Number: 0005886
>
> Gusshausstr. 27-29, room CA-05-38
> 1040 Vienna, Austria
> Phone: +43 (0)1 588 01 37688
> Fax: +43 (0)1 588 01 937688
> Email: melik-merkumians@xxxxxxxxxxxxxxxxx
> http://www.acin.tuwien.ac.at/
> --------------------------------------------------
>
> Von: 4diac-dev-bounces@xxxxxxxxxxx <4diac-dev-bounces@xxxxxxxxxxx> Im Auftrag von Sándor Bácsi
> Gesendet: Donnerstag, 4. Juni 2020 09:25
> An: 4diac developer discussions <4diac-dev@xxxxxxxxxxx>
> Betreff: [4diac-dev] Detecting self-cycles in ECC
>
> Hi,
>
> As a first step, I started to implement OCL constraints to detect inconsistencies in ECC. Such an inconsistency can be, for example, if we create self-cycles in the ECC. A self-cycle exists when there is a transition from an EC state back to itself. If there is no event associated with such a transition, it can happen that the executor may enter an infinite loop. Therefore, I think we should warn users about self-cycles gated by "1" or "[None]". I created an OCL invariant which can detect self-cycles, which are gated by "1" or "[None]".
>
> Gated by "[None]":
>
>
>    Gated by "1":
>
>
>
> Since I am a newbie to ECC, I would like to ask your opinion on this constraint. What do you think is this the right direction for detecting self-cycles? What previous experiences do you have with self-cycles in ECC?
>
> Cheers,
> Sándor
>
> _______________________________________________
> 4diac-dev mailing list
> 4diac-dev@xxxxxxxxxxx
> To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/4diac-dev