Show all possible traces for a given petri net

Isabel

Hello all,

might be a trivial question, but I have not yet found any solution...

Is there an automatism to derive all possible traces which can be generated by a given (not very complex) petri net? I have discovered a few simple models for which I want to know the possible traces. Because the models do not contain loops, it should be a finite set. (However loops could also be expressed by e.g. using a regular expression like asteriks (AB*C = ABC, ABBC, ABBBC, ABBBBC, ....). By looking at the set of possible traces I could verify whether the model is over- or undergeneralizing.

So far I have looked into transition systems (hard to read) and also the unfolding plugins available in ProM 6. Neither has given me the desired information. Knowing which traces can in theory be generated by a specific petri net would support people not acquainted with petri nets and creates a better understanding of the model behaviour (at least for not very complex petri nets, e.g. with a dozen possible traces).

Would it in principle be feasible to derive all possible traces in a petri net using a combined depth-first search / repeated token game approach? Sorry if this is a really stupid question...

Does anybody know how to handle this or can anyone point out some keywords which I could use for further research towards a solution?

Thank you very much in advance,
Isabel

aadrians

Hi Isabel,

To my knowledge, there is no ProM 6 plugins that generate possible traces for a given Petri net. The best thing that I can come up with is the transition system of petri net, which you already tried.

If you want to measure how over/under generalize an event log to a process model (being precision and generalization conformance metrics), try the "Measure precision/generalization" plug-in in the PNetAlignmentAnalysis package of ProM 6. Before you use the plugin, run the "Replay a Log on Petri Net for Conformance Analysis" plug-in to get a "PNRepResult" object that it is required for the former plug-in. The measurement is based on the one mentioned in [1].

Regards,

Arya

Reference:

[1] van der Aalst, W.M.P., Adriansyah, A., & Dongen, B.F. van. Replaying History on Process Models for Conformance Checking and Performance Analysis. WIREs Data Mining Knowl Discov 2012, 2: 182-192. doi: 10.1002/widm.1045

Isabel

Hello Arya,

thank you very much for your answer. I will try what you suggested and let you know how it works out.

Anyway, I'm still open for further input - maybe other software/tools, even articles describing an algorithmic approach to the problem, keywords or such...

Best wishes,
Isabel

aadrians

Hi Isabel, 

To generate exhaustive traces of a given net, you can simulate the net using the CPN Tools (http://cpntools.org/). I have an informal post on my blog on how to generate event logs from a CPN model (http://blog.adriansyah.info/archives/295). Feel free to use it.

Regards,

Arya

likewinthomas

HI ARYA

Greeting of the day!

Can Process mining be applied in BIO INFORMATICS?

if yes can you please let me know how and where?

else can you please tell me why?

Thank you sir, Hoping for a needful reply at the soonest

Duanchenchen

aadrians said:

Hi Isabel, 

To generate exhaustive traces of a given net, you can simulate the net using the CPN Tools (http://cpntools.org/). I have an informal post on my blog on how to generate event logs from a CPN model (http://blog.adriansyah.info/archives/295). Feel free to use it.

Regards,

Arya

Hi aadrians

I really want to learn how to generate event logs from a CPN tools. But your blog link has expired. Can you share a link to your latest blog to me?

Thank you sir, Hoping for a needful reply at the soonest

Regards,

Duan