Semantics-based business process model similarity – scrapbooking gas dryer vs electric dryer


Business process modeling has become an accepted means for designing and describing business operations. As a result, comparing and aligning business process models within and between organizations is increasingly important. However, due to differing use of modeling languages and domain languages for labeling models and their elements, model comparison is a non-trivial task. Presently, it is to be performed manually. For easing this workload, we present a novel approach for determining semantic similarity in an automated manner, directed at supporting business analysis through semantic reasoning. Keywords: Business process modeling, model comparison, semantic similarity gas examples, semantic reasoning

2]. For establishing electronic business, the underlying processes, required information and subsequent IT-support need to be described precisely. Over the past decades, business process modeling has become an accepted means for designing and describing business operations in enterprises within and across company boundaries. Such models describe interrelated business objects and business activities in a specific sequence, expressed in a certain modeling language with elements labeled in natural language. Thereby, major tasks in working with models are the analysis of these descriptions for the purpose of quality and compliance checking and detecting commonalities with models of different srcin [3, 4]. This involves checking models with regard to the modeling languages and, most importantly, the domain language used for labeling the model elements. If the choice of words of the labels representing the business semantics is not gasbuddy trip dominated by rules, models are semantically heterogeneous, not only concerning their modeling language, but more importantly, concerning their domain language. This makes their comparison or integration a non-trivial task [5, 6].

As a result, differing types of models and dissimilarly applied business terminology prevent direct automated business process interactions without p gasol prior manual preparation efforts for resolving discrepancies. Typically, these challenges arise in all kinds of e-business integration projects, such as enterprise architecture, data and process integration scenarios [7]. Especially at the time of mergers and acquisitions and setting-up business collaborations where the models to be integrated srcinate from different independent sources, semantic analysis requires extensive intellectual efforts and time [8]. Therefore, in practice the problem presently to be tackled is the task of having to analyze hundreds of business process models manually. Models need to be compared gas in dogs regarding the intended meaning of their elements and their structure, whereas structural analysis cannot be performed until successful alignment of the domain language [9]. For easing this workload, automated support is deemed desirable by way of enabling (semi-)automatic alignment of process models concerning their semantic similarity. For supporting the analysis of models with regard to the intended meaning of language concepts present, the idea of applying semantic technologies in business process modeling has been suggested [10]. Semantic annotation of business process models has been proposed to allow for analyzing and comparing models [11, 6, 12]. As a complement to these efforts, we here present our method of analyzing and reasoning over business process models based on the domain facts contained. We report on our research and continue with presenting a typical application scenario encountered, followed by a description of our method based on Semantic Web technologies and show its application. We conclude with an evaluation and discussion of our proposition together with a view onto related work and future research directions.

To demonstrate our method, we show a typical application scenario for business process model integration. The motivating example is a situation occurring within enterprises at the time of outsourcing. It is a case of a travel agency implant where a company decided to hand over the booking of travel related services to an agency belonging to an independent business partner but installed on the premises. Fig. 1 shows two example models 76 gas station hours, an Event-

Cancel bookingsSend reservation problem notificationVReservations undoneCustomer wishes to continuePayment processedNotify customerTravel reservation successfulget travel datacheck credit cardhandle invalid credit cardsend invoicehotel reservationflight reservationcar reservationgenerate confirmation reservation successfulnotify customercreate invoicesend emailall reservations successfulprepare emailcredit card invalid

compare the types of modeling languages and language elements based on their meta-models. For example, an EPC is more similar to a UML activity model than to a UML class model. This is because EPC and UML activity models are both expressed in business process modeling languages, whereas UML class models are expressed in a structure modeling language. Analogously, on the element level, an EPC function is more similar gas engine tom to a UML activity than to a UML class. We have implemented a simple similarity metric, namely