DOLCE and Pi-Calculus Rendezvous Semantics for Business Processes Violeta Damjanović Salzburg Research, Austria 01-05 June, 2008 1 st International Workshop on Knowledge Reuse and Reengineering over the Semantic Web (KRRSW 2008) hosted by the 5 th ESWC 200 8 , Costa Adeje , Tenerife, Canary Islands

Problem Description ImportNET real business domains Staticity of ontological models vs. dynamicity of business processes Semantic reengineering of the static knowledge domains to support dynamic business processes OMG’ ODM standard for model driven ontology development OMG’ MOF that defines the metadata architecture for MDA

ImportNET real business domains

Staticity of ontological models vs. dynamicity of business processes

Semantic reengineering of the static knowledge domains to support dynamic business processes

OMG’ ODM standard for model driven ontology development

OMG’ MOF that defines the metadata architecture for MDA

Approaches to Business Modeling Abstract framework to business process specification Functional view – Behavioral view – Structural view Functional specification based on IDEF (Integration DEFinition language) Well formalized; standardized by NIST; resulting specification is too complex Process specification using EPC Not formally defined; Syntax and semantics is not precise enough OO approach to structural modeling Objects + Classes + Relationships (association, aggregation, generalization) Using UML for business processes Semi-formal method; the semantics is not precisely defined Meta-model of business processes The idea is to combine formal methods with meta-models

Abstract framework to business process specification

Functional view – Behavioral view – Structural view

Functional specification based on IDEF (Integration DEFinition language)

Well formalized; standardized by NIST; resulting specification is too complex

Process specification using EPC

Not formally defined; Syntax and semantics is not precise enough

OO approach to structural modeling

Objects + Classes + Relationships (association, aggregation, generalization)

Using UML for business processes

Semi-formal method; the semantics is not precisely defined

Meta-model of business processes

The idea is to combine formal methods with meta-models

Business Process Transformation Trinity DDPO – OWL BPEL, WSDL Pi-Calculus Ontological models Business Processes Process theory

DDPO – OWL

BPEL, WSDL

Pi-Calculus

DDPO DDPO (DOLCE D&S Plan & Task Ontology) DDPO theoretical model: to help us getting a good understanding of business models at the different levels of abstraction and to provide implicit rules for expressing the facts that explain behavior and structure of the abstract business processes

DDPO (DOLCE D&S Plan & Task Ontology)

DDPO theoretical model:

to help us getting a good understanding of business models at the different levels of abstraction and

to provide implicit rules for expressing the facts that explain behavior and structure of the abstract business processes

BPEL, WSDL BPEL enables realization of SOA through composition, coordination and orchestration of WS BPEL syntax is defined by a BPEL XMLS which describes the BPEL basic activities, partner activities and structural activities

BPEL enables realization of SOA through composition, coordination and orchestration of WS

BPEL syntax is defined by a BPEL XMLS which describes the BPEL basic activities, partner activities and structural activities

What is the Pi-Calculus? mathematical formalisms for describing and analyzing properties of concurrent computation developed by Robin Milner in the 1990s CCS, CSP, ACP… widely used in AI (Artificial Intelligence) major areas that use the Pi-calculus: ERLANG language (Robin Milner – consultant to ATT and British Telephone) LOTOS (very complex and delicate temporal reasoning problems in the NASA space missions) a foundation for the methodologies for BPM (Business Process Modeling) Pi-C alculus

What is the Pi-Calculus?

mathematical formalisms for describing and analyzing properties of concurrent computation developed by Robin Milner in the 1990s

CCS, CSP, ACP…

widely used in AI (Artificial Intelligence)

major areas that use the Pi-calculus:

ERLANG language (Robin Milner – consultant to ATT and British Telephone)

LOTOS (very complex and delicate temporal reasoning problems in the NASA space missions)

a foundation for the methodologies for BPM (Business Process Modeling)

The core syntax of Pi- C alculus

P rocesses are written using the following syntax (formal model): summation parallel composition process expression process definition transition

P rocesses are written using the following syntax (formal model):

Mapping the DDPO Elementary task to the Pi-Calculus Operator 1. transforming the DDPO Elementary task An elementary task is an atomic task ElementaryTask(x) = df  y. Component(x,y)  Task(y)

1. transforming the DDPO Elementary task

An elementary task is an atomic task

ElementaryTask(x) = df  y. Component(x,y)  Task(y)

2. transforming the DDPO Component A component relation is a proper part relation qualified by a description in which the proper part are involved Component(x,y) = df ProperPart(x,y)   d,z,w. Description(d)  Role(z)  Role(w)  Uses(d,z)  Uses(d,w)  Selects(z,x)  Selects(w,y)

2. transforming the DDPO Component

A component relation is a proper part relation qualified by a description in which the proper part are involved

Component(x,y) = df ProperPart(x,y)   d,z,w. Description(d)  Role(z)  Role(w)  Uses(d,z)  Uses(d,w)  Selects(z,x)  Selects(w,y)

3. transforming the DPPO Task A task is a course defined by a plan in which at least one intentional agentive role or intentional figure has a desire attitude towards task Task(x) = df Course(x)   y,z. Plan(y)   Defines(y,x)   ((IntentionalAgentiveRole(z)  IntentionalFigure(z))   Uses(y,z)  DesireTowards(z,x)

3. transforming the DPPO Task

A task is a course defined by a plan in which at least one intentional agentive role or intentional figure has a desire attitude towards task

Task(x) = df Course(x)   y,z. Plan(y)   Defines(y,x)   ((IntentionalAgentiveRole(z)  IntentionalFigure(z))   Uses(y,z)  DesireTowards(z,x)

OWL2BPEL Metamodels and Transformation Models The main objectives of the MOF specification To support multiple metamodels and models, and To enable their extensibility, integration, and generic model and metamodel management Mapping the source model to the target model Source model : Ontology based on DDPO model Target model : BPEL process that can be deployed to a standalone BPEL workflow engine

The main objectives of the MOF specification

To support multiple metamodels and models, and

To enable their extensibility, integration, and generic model and metamodel management

Mapping the source model to the target model

Source model : Ontology based on DDPO model

Target model : BPEL process that can be deployed to a standalone BPEL workflow engine

Transformation scenario

Transformation scenario

Transformation plan

Transformation plan

Summary and Conclusion ImportNET is trying to bring the Semantic Web technologies to mechatronic engineering Mechatronic engineering is multi disciplinary and therefore needs multiple ontologies The engineering process needs process modeling The engineering artifacts need structural modeling Therefore our ontologies need to capture the dynamic and the static aspects of the domains

ImportNET is trying to bring the Semantic Web technologies to mechatronic engineering

Mechatronic engineering is multi disciplinary and therefore needs multiple ontologies

The engineering process needs process modeling

The engineering artifacts need structural modeling

Therefore our ontologies need to capture the dynamic and the static aspects of the domains

Thank you ! Questions ? Contact : violeta.damjanovic @ salzburgresearch.at http://www.salzburgresearch.at

Thank you !

Questions ?

Contact :

violeta.damjanovic @ salzburgresearch.at

http://www.salzburgresearch.at