Discovering Stochastic Petri Nets with Arbitrary Delay Distributions from Event Logs
Capturing the performance of a system or business process as accurately as possible is important, as models enriched with performance information provide valuable input for analysis, operational support, and prediction. Due to their computationally nice properties, memoryless models such as exponentially distributed stochastic Petri nets have earned much attention in research and industry. However, there are cases when the memoryless property is clearly not able to capture process behavior, e.g., when dealing with fixed time-outs.
We want to allow models to have generally distributed durations to be able to capture the behavior of the environment and resources as accurately as possible. For these more expressive process models, the execution policy has to be specified in more detail. In this paper, we present and evaluate process discovery algorithms for each of the execution policies. The introduced approach uses raw event execution data to discover various classes of stochastic Petri nets. The algorithms are based on the notion of alignments and have been implemented as a plug-in in the process mining framework ProM.
KeywordsProcess mining Stochastic Petri nets Generally distributed transitions
- 4.Leclercq, E., Lefebvre, D., Ould El Mehdi, S.: Identification of timed stochastic petri net models with normal distributions of firing periods. In: Information Control Problems in Manufacturing, vol. 13, pp. 948–953 (2009)Google Scholar
- 5.Anastasiou, N., Horng, T., Knottenbelt, W.: Deriving generalised stochastic Petri Net performance models from high-precision location tracking data. In: VALUETOOLS’11, ICST, pp. 91–100 (2011)Google Scholar
- 8.Adriansyah, A., van Dongen, B., van der Aalst, W.: Conformance checking using cost-based fitness analysis. In: EDOC 2011, pp. 55–64. IEEE (2011)Google Scholar
- 12.Fahland, D., van der Aalst, W.M.P.: Repairing process models to reflect reality. In: Barros, A., Gal, A., Kindler, E. (eds.) BPM 2012. LNCS, vol. 7481, pp. 229–245. Springer, Heidelberg (2012) Google Scholar
- 13.Buijs, J.C.A.M., La Rosa, M., Reijers, H.A., van Dongen, B.F., van der Aalst, W.M.P.: Improving business process models using observed behavior. In: Cudre-Mauroux, P., Ceravolo, P., Gašević, D. (eds.) SIMPDA 2012. LNBIP, vol. 162, pp. 44–59. Springer, Heidelberg (2013) Google Scholar
- 15.Kooperberg, C., Stone, C.J.: Logspline density estimation for censored data. J. Comput. Graph. Stat. 1(4), 301–328 (1992)Google Scholar
- 16.Buchholz, R., Krull, C., Horton, G.: Reconstructing model parameters in partially-observable discrete stochastic systems. In: Al-Begain, K., Balsamo, S., Fiems, D., Marin, A. (eds.) ASMTA 2011. LNCS, vol. 6751, pp. 159–174. Springer, Heidelberg (2011)Google Scholar
- 17.Wombacher, A., Iacob, M.E.: Start time and duration distribution estimation in semi-structured processes. Technical report, Centre for Telematics and Information Technology, University of Twente (2012)Google Scholar