Publication Details
Precise Parameter Synthesis for Generalised Stochastic Petri Nets with Interval Parameters
Češka Milan, prof. RNDr., CSc. (DITS FIT BUT)
Paoletti Nicola (UOx)
parameter synthesis
Stochastic Petri Nets
continuous-time Markov Chains
tmeporal logic
We consider the problem of synthesising parameters affecting transition rates and probabilities in generalised Stochastic Petri Nets (GSPNs). Given a time-bounded property expressed as a probabilisitic temporal logic formula, our method allows computing the parameters values for which the probability of satisfying the property meets a given bound, or is optimised. We develop algorithms based on reducing the parameter synthesis problem for GSPNs to the corresponding problem for continuous-time Markov Chains (CTMCs), for which we can leverage existing synthesis algorithms, while retaining the modelling capabilities and expressive power of GSPNs. We evaluate the usefulness of our approach by synthesising parameters for two case studies.
@INPROCEEDINGS{FITPUB11302, author = "Milan \v{C}e\v{s}ka and Milan \v{C}e\v{s}ka and Nicola Paoletti", title = "Precise Parameter Synthesis for Generalised Stochastic Petri Nets with Interval Parameters", pages = "38--46", booktitle = "Proceedings of 16th International Conference on Computer Aided Systems Theory", series = "LNCS volume 10672", year = 2017, location = "Heidelberg, DE", publisher = "Springer Verlag", ISBN = "978-3-319-74726-2", doi = "10.1007/978-3-319-74727-9\_5", language = "english", url = "https://www.fit.vut.cz/research/publication/11302" }