Information Technology Reference
In-Depth Information
Their procedure is based on transformation rules that are applied directly to
a BP model. However, no data anomalies are described nor detected in their
approach. Besides, our use of well-known Petri nets algorithms makes it more
unlikely to introduce errors while implementing the procedure. In [11] the oppo-
site procedure is introduced, i.e. an approach for generating a BP model from
OLCs of different data objects is described.
Data anomalies in BP models have been addressed by several researchers.
Sadiq et al. [3] explain the importance of managing the data requirements in
BPs and introduce some ideas related to the modelling and validation of data,
such as the importance of considering the type of data and their structure. They
also state some data anomalies that may appear in a BP model, which in turn are
referenced by the authors in [4]. In that work, Sun et al. divide the same problems
into three main groups with one or more scenarios, and then they explain the
matching of every scenario with the data anomalies in [3]. Awad et al. describe
three kinds of data anomalies that can also be mapped to anomalies defined in the
previously mentioned work [2]. They have developed an approach for diagnosing
and automatically repairing these three kinds of problems on the basis of Petri
nets. A prototype has been implemented in Oryx [12]. Besides, they propose
some validation algorithms targeted at fixing these data anomalies, which are
being implemented to correct BPMN models. In our BP2OLC procedure, we use
the transformations described in that work to carry out step 1 of the BP2OLC
procedure. However, the mentioned work on data anomalies does not consider
the generation of OLCs from a BP model.
Finally, Sakr et al. have developed a framework for querying both control flow
and data flow perspectives of BPs [13]. Data perspective can be queried from
OLCs. However, no automatic generation of OLCs is included and the framework
is not targeted at the detection and management of data anomalies.
6 Conclusions and Future Work
In this paper we introduce a model-driven approach for the automatic generation
of a data-centered view of a BP composed of the life cycles of the data objects
the BP model has. It consists of mapping a BPMN model into a target semantic
domain, Petri nets, which allows us to use techniques specific to that domain,
in particular obtaining its reachability graph, for analysing the source model.
Then, the reachability graph is mapped into an OLC model. An advantage of
our procedure is that the resulting OLCs include information about the activ-
ities that are executed in each transition and, hence, it provides the same full
information required to understand BP execution as activity-centered process
diagrams.
Besides, our procedure is robust in the sense that it provides an accurate
result despite having a BP with data anomalies as input. Furthermore, we detail
how this procedure can be used to detect two kinds of data anomalies present in
a BP model. For each group of data anomalies identified, the following questions
have been answered: (i) what does the anomaloussituationmeanintermsofthe
