#### DMCA

## Quantifying Process Equivalence Based on Observed Behavior

### Cached

### Download Links

- [is.tm.tue.nl]
- [is.tm.tue.nl]
- [wwwis.win.tue.nl]
- [wwwis.win.tue.nl]
- [wwwis.win.tue.nl]
- DBLP

### Other Repositories/Bibliography

Citations: | 23 - 0 self |

### Citations

1577 |
A Calculus of Communicating Systems
- Milner
- 1982
(Show Context)
Citation Context ... these models have semantics. Moreover, we need to assume some equivalence notion (When are two models the same?). People working on formal methods have proposed a wide variety of equivalence notions =-=[1, 19, 26]-=-, e.g., two models may be identical under trace equivalence but are different when considering stronger notions of equivalence (e.g., bisimulation). Unfortunately, most equivalence notions provide a “... |

978 |
Coloured Petri Nets. Basic Concepts, Analysis Methods and Practical Use
- Jensen
- 1997
(Show Context)
Citation Context ...cies: 40: ABDE 85: ACDE 15: ADBE 20: ADCE (f) Fig. 1. Running example. p5 B C A E p1 p2 p3 To clarify the problem, let us consider Figure 1 where four process models (expressed in terms of Petri nets =-=[14, 22, 29]-=-) are depicted. These models dep4 p6 D p5 Fsscribe the booking of a trip, see the legend for the interpretation of the various transitions in the Petri nets, e.g., C refers to the booking of a flight.... |

382 | Workflow Mining: Discovering Process Models from Event Logs
- Aalst, Weijters, et al.
(Show Context)
Citation Context ...stract BPEL processes in the context of web services [9], choreography descriptions using WSCDL [23], or other ad-hoc notations [33]);sand/or process models discovered using process mining techniques =-=[5, 6]-=-. Given the co-existence of different models and different types of models, it is interesting to be able to compare process models. This comparison of process models applies to different levels rangin... |

317 | W.: Branching Time and Abstraction in Bisimulation Semantics
- Glabbeek, Weijland
- 1996
(Show Context)
Citation Context ... these models have semantics. Moreover, we need to assume some equivalence notion (When are two models the same?). People working on formal methods have proposed a wide variety of equivalence notions =-=[1, 19, 26]-=-, e.g., two models may be identical under trace equivalence but are different when considering stronger notions of equivalence (e.g., bisimulation). Unfortunately, most equivalence notions provide a “... |

309 | Discovering Models of Software Processes from EventBased Data - Cook, Wolf - 1998 |

289 |
Workflow Mining: A Survey of Issues and Approaches
- Aalst, Dongen, et al.
(Show Context)
Citation Context ...stract BPEL processes in the context of web services [9], choreography descriptions using WSCDL [23], or other ad-hoc notations [33]);sand/or process models discovered using process mining techniques =-=[5, 6]-=-. Given the co-existence of different models and different types of models, it is interesting to be able to compare process models. This comparison of process models applies to different levels rangin... |

253 | Mining Process Models from Workflow Logs - Agrawal, Gunopulos, et al. - 1998 |

183 | Inheritance of workflows: An approach to tackling problems related to change.” Theoret
- Aalst, Basten
(Show Context)
Citation Context ... these models have semantics. Moreover, we need to assume some equivalence notion (When are two models the same?). People working on formal methods have proposed a wide variety of equivalence notions =-=[1, 19, 26]-=-, e.g., two models may be identical under trace equivalence but are different when considering stronger notions of equivalence (e.g., bisimulation). Unfortunately, most equivalence notions provide a “... |

120 |
Free Choice Petri Nets, volume 40 of Cambridge Tracts
- Desel, Esparza
(Show Context)
Citation Context ...cies: 40: ABDE 85: ACDE 15: ADBE 20: ADCE (f) Fig. 1. Running example. p5 B C A E p1 p2 p3 To clarify the problem, let us consider Figure 1 where four process models (expressed in terms of Petri nets =-=[14, 22, 29]-=-) are depicted. These models dep4 p6 D p5 Fsscribe the booking of a trip, see the legend for the interpretation of the various transitions in the Petri nets, e.g., C refers to the booking of a flight.... |

78 |
Introduction to Evolutionary Computing. Natural Computing Series
- Eiben, Smith
- 2003
(Show Context)
Citation Context ...lems. Genetic process mining is the application of genetic algorithms to process mining. Genetic algorithms are an adaptive search technique that mimics the process of evolution in biological systems =-=[18]-=-. Figure 7 depicts the main steps of our genetic approach. Once the log is read (Step I), the algorithm randomly builds an initial population with a number of individuals (Step II). Every individual i... |

73 | A Machine Learning Approach to Workflow Management - Herbst - 2000 |

72 | Metrics for labeled Markov processes
- Desharnais, Gupta, et al.
(Show Context)
Citation Context ...o far, aim at a “true/false” answer. Moreover, they do not take into account that some parts of the process may be more important than others. Few people (e.g., Prakash Panangaden and Jose Desharnais =-=[15]-=-) have been working on probabilistic bisimulation using labeled Markov processes rather than labeled transition systems. See [15] for an excellent overview of this work and also links to the probabili... |

65 |
der Aalst, The ProM framework: A new era in process mining tool support
- Dongen, Medeiros, et al.
- 2005
(Show Context)
Citation Context ...on 5 we present the core results of this paper. We will show that we can define precision and recall measures using event logs containing typical behavior. These notions have been implemented in ProM =-=[16]-=-. In sections 6 and 7 we reason about what can be inferred based on the results of the precision and recall measures over different models with respect to a same given log. Finally, we discuss the app... |

53 | Genetic Process Mining
- Aalst, Medeiros, et al.
- 2005
(Show Context)
Citation Context ...ded in event logs as an aid for comparison. This makes the work quite different from the references mentioned in this section. Moreover, we show that this can be used in the context of process mining =-=[3, 5, 6]-=-.s3 Preliminaries This section introduces some of the basic mathematical and Petri-net related concepts used in the remainder. 3.1 Multi-sets, Sequences, and Matrices Let A be a set. IB(A) = A → IN is... |

39 |
C.: A survey of equivalence notions for net based systems
- Pomello, Rozenberg, et al.
- 1992
(Show Context)
Citation Context ...focus on the dynamics of the model and not on the syntactical structure (e.g., trace equivalence and bisimulation [1, 19, 26]). This paper uses Petri nets as a theoretical foundation [14, 22, 29]. In =-=[28]-=- an overview is given of equivalence notions in the context of Petri nets. See also [10] for more discussions on equivalence in the context of nets. Most authors translate a Petri net to a transition ... |

33 | Conformance testing: Measuring the fit and appropriateness of event logs and process models
- Rozinat, Aalst
- 2006
(Show Context)
Citation Context ...occur multiple times.sBefore comparing two process models using an event log, we first define the notion of fitness. This notion is inspired by earlier work on genetic mining and conformance checking =-=[25, 31]-=-. Definition 7 (Fitness). Let (N, M) be a marked Petri net and let L ∈ IB(T ∗ ) be a multi-set over T . 4 fitness((N, M), L) = ( � L(σ) |{i ∈ {0, |σ| − 1} | σ(i + 1) ∈ enabled(N, M, hd(σ, i))}| )/|L| ... |

31 | Process equivalence: Comparing two process models based on observed behavior
- Aalst, Medeiros, et al.
- 2006
(Show Context)
Citation Context ...sible according the the first model’s behavior. Recall measures are used to quantify how much of the first model’s behavior is covered by the second model. This paper extends the results presented in =-=[4]-=- in three ways. Firstly, we investigate what can be inferred about the values of the defined precision and recall measures for two models - say model 1 and model 3 - and some typical behavior - say lo... |

24 |
de Medeiros. Genetic Process Mining
- Alves
- 2006
(Show Context)
Citation Context ...ation system, but where health-care professionals determine the “careflow”. A variety of process mining algorithms have developed [5–7, 12, 21], including our approach based on genetic process mining =-=[3, 25, 8]-=-.sFig. 5. Example illustrating why all the tasks in the process models should be in the log. In this case, C and E are not in the log. Fig. 6. Example why the log should contain enough behavior such t... |

18 |
A behavioural pseudometric for metric labelled transition systems
- Breugel
(Show Context)
Citation Context ...two models. Also related is the work on metric labeled transition systems where the “behavioral difference” between states is a non-negative real number indicating the similarity between those states =-=[11]-=-. This way one can define a behavioral pseudometric to compare transition systems as shown in [11]. Note that this approach very much depends on an explicit notion of states and it is not clear how th... |

18 | Discovering workflow models from activities' lifespans - Pinter, Golani - 2004 |

16 |
Validation of process models by construction of process nets
- Desel
- 2000
(Show Context)
Citation Context ...uthors translate a Petri net to a transition system to give it semantics. However, there are also authors that emphasize the true-concurrency aspects when giving Petrisnets semantics. For example, in =-=[13]-=- the well-known concept of occurrence nets (also named runs) are used to reason about the semantics of Petri nets. Any model with formal/executable semantics (including Petri nets) can be translated t... |

11 |
der Aalst. Genetic process mining : a basic approach and its challenges
- Medeiros, Weijters, et al.
(Show Context)
Citation Context ...occur multiple times.sBefore comparing two process models using an event log, we first define the notion of fitness. This notion is inspired by earlier work on genetic mining and conformance checking =-=[25, 31]-=-. Definition 7 (Fitness). Let (N, M) be a marked Petri net and let L ∈ IB(T ∗ ) be a multi-set over T . 4 fitness((N, M), L) = ( � L(σ) |{i ∈ {0, |σ| − 1} | σ(i + 1) ∈ enabled(N, M, hd(σ, i))}| )/|L| ... |

10 |
Some equivalence results for free choice nets and simple nets and on the periodicity of live free choice nets
- Best, Shields
- 1983
(Show Context)
Citation Context ...ivalence and bisimulation [1, 19, 26]). This paper uses Petri nets as a theoretical foundation [14, 22, 29]. In [28] an overview is given of equivalence notions in the context of Petri nets. See also =-=[10]-=- for more discussions on equivalence in the context of nets. Most authors translate a Petri net to a transition system to give it semantics. However, there are also authors that emphasize the true-con... |

10 | Finding Trading Partners to Establish Ad-Hoc Business
- Wombacher, Mahleko
(Show Context)
Citation Context ...ptions (e.g., the Partner Interface Processes in RosettaNet [30], the abstract BPEL processes in the context of web services [9], choreography descriptions using WSCDL [23], or other ad-hoc notations =-=[33]-=-);sand/or process models discovered using process mining techniques [5, 6]. Given the co-existence of different models and different types of models, it is interesting to be able to compare process mo... |

4 | An Efficient Algorithm for Branching and Stuttering Equivalence - Groote, Vaandrager - 1990 |