Semantic Annotation for Process Models: - Department of Computer ...
Semantic Annotation for Process Models: - Department of Computer ... Semantic Annotation for Process Models: - Department of Computer ...
38 CHAPTER 3. STATE OF THE ART modeling language for execution languages, such as BPEL4WS [58] and BPML [11]. The BPMN specification provides a mapping between the graphics of the notation and the constructs of those formal languages. BPMN intends to provide businesses with the capability of understanding their internal business procedures in graphical notations and give organizations the ability to communicate these procedures in a standard manner. BPMN is initiated as a standard process modeling language for conventional business, B2B and services process modeling. Hence BPMN has the capabilities of handling B2B business process concepts, such as public and private processes and choreograhies, as well as advanced modeling concepts, such as exception handling and transaction compensation in addition to the traditional business process notations [12]. Private business processes are those internal processes to a organization and they are traditional business processes or workflows. Public processes are also called abstract processes or interface processes according to the BPMI terminology. They represents the interactions between a private business process and another process or participant. Choreographies represents collaboration processes. A collaboration process depicts a sequence of activities that represent the messages being sent between the entries involved. According to the BPMI terminology, the modeling constructs are the elements with corresponding notations. The main elements of BPMN Version 1.0 [12] are event, task, process/sub-process, sequence flow, message flow, pool, lanes, data object, fork (ANDsplit), join (AND-join), decision/branching point (OR-split), merging (OR-join), looping, etc. We refer those elements to the perspectives in the BPM systems paradigm. A sub-process is a compound activity included within a process. The collapsed/expanded sub-process can hide and show the details of the sub-process. Hereby, the structural perspective can be represented by BPMN through such a mechanism. Respectively, task, process/sub-process are elements for the operational/functional perspective. Tasks and processes are controlled through event, sequence flow, fork, join, decision, merging and looping. Resource can be represented by data object. Organizations are usually represented by lanes in the pool. The data transaction is specified through message flow together with data object. The details of the BPMN specifications from BPMI can refer Appendix A. The names and the notations of the BPMN elements used in this section are from the standard proposal. When a modeling tool implements a certain modeling language (e.g. BPMN), the names and the notations might be slightly changed to adjust the tool implementation. For example, task, process/subprocess are implemented as "Logical Process" in an enterprise modeling tool Metis [183]. 3.1.6 Categorizing the modeling constructs of process modeling languages The investigation of existing process modeling languages has shown the diversity of modeling constructs. The modeling constructs of the process modeling languages is categorized in Table 3.1 according to the six process perspectives defined in the paradigm of BPM systems.
3.2. SEMANTIC INTEROPERABILITY AND PROCESS ONTOLOGIES 39 3.2 Semantic Interoperability and Process Ontologies Semantic heterogeneity baffles the effective management of distributed knowledge, which relates to semantic interoperability under an extensive system exchange and integration situation. Semantic interoperability issue has been studied in different domains and applications such as schema and data integration of databases, meta-data interoperation among distributed digital libraries , agent-based Web services discovery and composition, enterprise integrations and so on. Research on semantic interoperability generally is categorized as: mapping and intermediary approach. The mapping approach can be sub-classified into point-to-point mapping and global mapping. Pointto-point mapping is the particular mapping built for two participating systems. Such a mapping can maximally preserve the original semantics of each system, but it is costly when a large number of systems need to interchange with each other or new unexpected system needs to participate. Global mapping attempts to construct mappings between participant systems and a global schema [24] [17] [66]. This requires to build a global schema which is designed to be dependent of particular schemas or applications [152]. The problem of this solution is not portable and does not adapt well to the addition of new systems. The intermediary approach is to make use of intermediary mechanisms to coordinate heterogeneous semantics. Domain-specific knowledge, mapping knowledge, or rules are modeled by the intermediary mechanisms such as mediators, agents, ontologies, etc. [132]. Generally, this approach uses a machine understandable definition of concepts and relationships between concepts so that there is a shared common understanding within a community. The knowledge or rules are domain specific, but independent of particular schemas and applications. However, one needs additional tools to actually capture and represent the knowledge (i.e., specifications and mappings) needed in order to resolve semantic conflicts. Semantic Web technology and some research results of ontology have initiated larger amount of research interests on the intermediary approach, especially applied in the applications of collaborative business (such as interoperation of enterprise processes, Web services, etc.). We apply the intermediary approach in this research work. Hence, the top-level ontology — process ontology for business process representation, and domain ontology for a given business domain are used as the intermediaries in our approach. The process ontology is used to reconcile the heterogeneous semantics of process modeling constructs (i.e. meta-model semantics) existing in different process modeling languages. It indicates that a process ontology should include a set of meta-concepts that are able to describe the semantics of process models. In this section we survey a number of process ontologies having different motivations: BWW ontology [204] for building a modeling fundamental of information systems, MIT process handbook [99] for organizing process knowledge, TOVE ontologies [31] for creating a set of ontologies to support enterpise modeling, PSL [120] for serving as an interlingua of all types of manufacturing processes, PIF [85] for exchanging business process models across different formats and schemas, OWL-S [198] for establishing a descriptive language of Web services, WSMO [209] for describing various aspects related to Semantic Web services, POP* [139] for providing a mapping mechanism between enterprise models and modeling tools, and UEML2 [140] for creating an intermediate language of various existing modeling languages. The process perspectives of the paradigm of BPM systems
- Page 8 and 9: iv CONTENTS 3 State of the Art 31 3
- Page 10 and 11: vi CONTENTS 7.5.2 Semantic reasonin
- Page 12 and 13: viii CONTENTS H PSAM Annotation Res
- Page 14 and 15: x LIST OF FIGURES 6.1 System module
- Page 16 and 17: xii LIST OF FIGURES
- Page 18 and 19: xiv LIST OF TABLES
- Page 20 and 21: xvi PREFACE My tremendous gratitude
- Page 23 and 24: Chapter 1 Introduction Business pro
- Page 25 and 26: 1.2. PROBLEM STATEMENT AND RESEARCH
- Page 27 and 28: 1.4. APPROACH AND SCOPE 7 1.4.1 Sem
- Page 29 and 30: 1.6. MAJOR CONTRIBUTIONS 9 1.6 Majo
- Page 31 and 32: Chapter 2 Problem Setting In this c
- Page 33 and 34: Figure 2.1: Zachman Enterprise Arch
- Page 35 and 36: 2.2. MODELING BASIS 15 meta-model o
- Page 37 and 38: 2.3. INFORMATION SYSTEMS AND SEMANT
- Page 39 and 40: 2.3. INFORMATION SYSTEMS AND SEMANT
- Page 41 and 42: 2.3. INFORMATION SYSTEMS AND SEMANT
- Page 43 and 44: 2.4. SEMANTIC INTEROPERABILITY 23 3
- Page 45 and 46: 2.5. BUSINESS PROCESS MODEL 25 2.5
- Page 47 and 48: 2.6. PROCESS KNOWLEDGE MANAGEMENT 2
- Page 49 and 50: 2.7. SUMMARY 29 • Creation or imp
- Page 51 and 52: Chapter 3 State of the Art This cha
- Page 53 and 54: 3.1. PROCESS MODELING LANGUAGES 33
- Page 55 and 56: 3.1. PROCESS MODELING LANGUAGES 35
- Page 57: Table 3.1: Modeling constructs of d
- Page 61 and 62: 3.2. SEMANTIC INTEROPERABILITY AND
- Page 63 and 64: 3.2. SEMANTIC INTEROPERABILITY AND
- Page 65 and 66: 3.2. SEMANTIC INTEROPERABILITY AND
- Page 67 and 68: Representation primitives Process P
- Page 69 and 70: 3.3. GOAL MODELING 49 From the surv
- Page 71 and 72: 3.4. SEMANTIC ANNOTATION METHODS AN
- Page 73 and 74: 3.4. SEMANTIC ANNOTATION METHODS AN
- Page 75 and 76: 3.4. SEMANTIC ANNOTATION METHODS AN
- Page 77 and 78: 3.4. SEMANTIC ANNOTATION METHODS AN
- Page 79 and 80: 3.6. SUMMARY 59 In the goal modelin
- Page 81: Part II Design and Application 61
- Page 84 and 85: 64 CHAPTER 4. SEMANTIC ANNOTATION F
- Page 86 and 87: 66 CHAPTER 4. SEMANTIC ANNOTATION F
- Page 88 and 89: 68 CHAPTER 4. SEMANTIC ANNOTATION F
- Page 90 and 91: 70 CHAPTER 4. SEMANTIC ANNOTATION F
- Page 92 and 93: 72 CHAPTER 4. SEMANTIC ANNOTATION F
- Page 94 and 95: 74 CHAPTER 4. SEMANTIC ANNOTATION F
- Page 96 and 97: 76 CHAPTER 4. SEMANTIC ANNOTATION F
- Page 98 and 99: 78 CHAPTER 4. SEMANTIC ANNOTATION F
- Page 100 and 101: 80 CHAPTER 4. SEMANTIC ANNOTATION F
- Page 102 and 103: 82 CHAPTER 4. SEMANTIC ANNOTATION F
- Page 104 and 105: 84 CHAPTER 5. GOAL ANNOTATION proce
- Page 106 and 107: 86 CHAPTER 5. GOAL ANNOTATION in a
3.2. SEMANTIC INTEROPERABILITY AND PROCESS ONTOLOGIES 39<br />
3.2 <strong>Semantic</strong> Interoperability and <strong>Process</strong> Ontologies<br />
<strong>Semantic</strong> heterogeneity baffles the effective management <strong>of</strong> distributed knowledge,<br />
which relates to semantic interoperability under an extensive system exchange and<br />
integration situation. <strong>Semantic</strong> interoperability issue has been studied in different domains<br />
and applications such as schema and data integration <strong>of</strong> databases, meta-data<br />
interoperation among distributed digital libraries , agent-based Web services discovery<br />
and composition, enterprise integrations and so on. Research on semantic interoperability<br />
generally is categorized as: mapping and intermediary approach. The mapping<br />
approach can be sub-classified into point-to-point mapping and global mapping. Pointto-point<br />
mapping is the particular mapping built <strong>for</strong> two participating systems. Such a<br />
mapping can maximally preserve the original semantics <strong>of</strong> each system, but it is costly<br />
when a large number <strong>of</strong> systems need to interchange with each other or new unexpected<br />
system needs to participate. Global mapping attempts to construct mappings between<br />
participant systems and a global schema [24] [17] [66]. This requires to build a global<br />
schema which is designed to be dependent <strong>of</strong> particular schemas or applications [152].<br />
The problem <strong>of</strong> this solution is not portable and does not adapt well to the addition <strong>of</strong><br />
new systems. The intermediary approach is to make use <strong>of</strong> intermediary mechanisms<br />
to coordinate heterogeneous semantics. Domain-specific knowledge, mapping knowledge,<br />
or rules are modeled by the intermediary mechanisms such as mediators, agents,<br />
ontologies, etc. [132]. Generally, this approach uses a machine understandable definition<br />
<strong>of</strong> concepts and relationships between concepts so that there is a shared common<br />
understanding within a community. The knowledge or rules are domain specific, but independent<br />
<strong>of</strong> particular schemas and applications. However, one needs additional tools<br />
to actually capture and represent the knowledge (i.e., specifications and mappings)<br />
needed in order to resolve semantic conflicts.<br />
<strong>Semantic</strong> Web technology and some research results <strong>of</strong> ontology have initiated larger<br />
amount <strong>of</strong> research interests on the intermediary approach, especially applied in the applications<br />
<strong>of</strong> collaborative business (such as interoperation <strong>of</strong> enterprise processes, Web<br />
services, etc.). We apply the intermediary approach in this research work. Hence, the<br />
top-level ontology — process ontology <strong>for</strong> business process representation, and domain<br />
ontology <strong>for</strong> a given business domain are used as the intermediaries in our approach.<br />
The process ontology is used to reconcile the heterogeneous semantics <strong>of</strong> process<br />
modeling constructs (i.e. meta-model semantics) existing in different process modeling<br />
languages. It indicates that a process ontology should include a set <strong>of</strong> meta-concepts<br />
that are able to describe the semantics <strong>of</strong> process models. In this section we survey<br />
a number <strong>of</strong> process ontologies having different motivations: BWW ontology [204] <strong>for</strong><br />
building a modeling fundamental <strong>of</strong> in<strong>for</strong>mation systems, MIT process handbook [99]<br />
<strong>for</strong> organizing process knowledge, TOVE ontologies [31] <strong>for</strong> creating a set <strong>of</strong> ontologies<br />
to support enterpise modeling, PSL [120] <strong>for</strong> serving as an interlingua <strong>of</strong> all types<br />
<strong>of</strong> manufacturing processes, PIF [85] <strong>for</strong> exchanging business process models across<br />
different <strong>for</strong>mats and schemas, OWL-S [198] <strong>for</strong> establishing a descriptive language<br />
<strong>of</strong> Web services, WSMO [209] <strong>for</strong> describing various aspects related to <strong>Semantic</strong> Web<br />
services, POP* [139] <strong>for</strong> providing a mapping mechanism between enterprise models<br />
and modeling tools, and UEML2 [140] <strong>for</strong> creating an intermediate language <strong>of</strong> various<br />
existing modeling languages. The process perspectives <strong>of</strong> the paradigm <strong>of</strong> BPM systems