• No results found

connected by the symbol “:-”. A rule body defines a set of conditions. A rule head represents the conclusion (new knowledge) inferred on condition (body) being true. Variables are defined by using the symbol “?” preceding the variable name. Con- straints are modeled using the symbol “!-” followed by the set of conditions (body). A constraint is violated when its condition or body is true and the target ontology becomes inconsistent. In the example ontology, a constraint is defined to ensure that students are not allowed to lead a project (see axiom leadsProjectConstraint in Listing 2.1).

WSML-Flight also allows for querying of ontologies. A query consists of a body alone, without the head. For example, the query supervises(?x, ?y, ?z) on the example ontology (see Listing 2.1) would return Christian, Mario and SUPER for the variables ?x, ?y and ?z, respectively. For further details on WSML-Flight syntax, we refer the reader to [dBFK+08, dBLPF06].

The reasoning tasks that can be performed with WSML-Flight are realized by Datalog queries [Dah97] and include: consistency check, entailment and retrieval [dBFK+08]. The consistency check of a WSML-Flight ontology is performed by querying for the empty clause. If the result set is empty, this means that the empty clause could not be derived and the queried ontology is satisfiable, otherwise it is not [dBFK+08]. This reasoning task is used in Chapter 5 of this thesis. The reason- ing task of entailment by a WSML-Flight ontology is performed by using queries that contain no variables. If the resulting set is non-empty, then the fact is entailed by the queried ontology, otherwise it is not [dBFK+08]. Entailment is used in Chap- ter 4 of this thesis. The retrieval is performed in a similar way to entailment, the difference being that the posed query contains free variables. The result set then contains all instances for which the query expression is entailed by the queried on- tology [dBFK+08]. Retrieval reasoning task is used in Chapters 4, 5 and 6 of this thesis.

2.4

Semantic Business Process Management

In this section we will draw a connection between both fields mentioned before, namely Business Process Management and semantic technologies. The research field of Semantic Business Process Management (SBPM) was born from a need to increase the level of automation within the BPM lifecycle. In the SBPM vision pa- per [HLD+05], the authors propose to combine the technologies developed in the Semantic Web services (SWS) community [SGA07] with BPM community efforts in order to create one consolidated technology called SBPM. The remainder of this sec- tion clarifies the aims and the approach of SBPM.

The collection of everything that matches the definition of a process from Section 2.1 in an enterprise can be referred to as the enterprise process space. Different roles are involved in BPM efforts, as illustrated in Fig. 2.7, but we can broadly identify two perspectives on BPM: i) the business perspective on enterprise operations and ii) the actual execution of the operations by a number of IT systems, resources, and human labor. We adopt the view of Hepp and Roman [HR07] who see BPM as an approach for managing the execution of IT-supported business operations from a managerial view rather than from a technical perspective.

Figure 2.7: The critical IT/process divide. Source: [HLD+05]

The main idea in SBPM is to semantically represent the two perspectives (busi- ness and IT) using ontology languages and to employ machine reasoning for the automated or semi-automated translation tasks between the two perspectives. In this way, it is possible to overcome the problem of disconnectedness of management and execution levels through SBPM. At present however, businesses lack a machine readable representation of their process space. Hepp et al. [HLD+05] analysis is that therefore “the business process space inside the organization as a whole is not ac- cessible on a semantic level”. Not having this is a major obstacle towards achieving higher level of automation in BPM. The authors also claim that the current research results in the area of Semantic Web and Semantic Web services (ontology languages, repositories, reasoners and query languages) should be applied to the BPM space in order to achieve the vision of SBPM. Hence, the core idea of the SBPM approach is to i) represent and semantically describe each existing atomic and composite process in an organization as a Semantic Web service, ii) express business expert needs as WSMO goals, iii) use an existing SWS execution environment to mediate between business goals and available processes.

Hepp et al. [HLD+05] formulate two general usage scenarios for SBPM: i) query- ing the process space, and ii) manipulating the process space. For instance, in the first scenario, a query could give an answer to the natural language question, “In which processes is resource x used?”, thereby exploiting both, the process space as well as further enterprise knowledge, i.e. resource structure. The second scenario subsumes the following tasks: creating a new business process, modify an existing process or abandon an outdated process. This scenario requires querying function- ality plus the possibility to actually invoke Web services and execute processes. The work presented in this thesis loosely relates to the scenario i) in Chapter 6, where we present an approach for querying of business processes.

SBPM has received growing attention through workshops and conference tracks organized on this topic, e.g. cf. [HHK+07, BHK+08]. Although SBPM also aims at formalizing the knowledge within the BPM domain, in our opinion it is too strongly driven by the SWS community which tries to apply the existing mechanisms of au- tomated discovery, composition, mediation [SGA07] of SWSs in the BPM domain.

2.5 Summary 29