Research On Mechanism For Semantics Of Software As A Service Based On Public Open Ontology

The rapid development of Internet has greatly changed the software development,and software systems are evolving towards socialized services through the network.As the typical representatives of these software systems,Web services have attracted attentions from both industrial and academic researchers.As the number of Web services grows and the diversity increases,automatically discovering,composing and invoking Web services in the open dynamic Internet environment has become an urgent issue.Semantic Web services have made some progress in augmenting the semantics of Web services and achieving awareness among different Web services.However,the following issues are still unsolved.Firstly,current semantic annotation methods rely heavily on manually built domain ontologies.Secondly,the heterogeneity problem among ontologies severely hampers the semantic interoperability among Web services that are annotated by utilizing different ontologies.Thirdly,how to achieve fast and accurate interoperability among Web services in the open dynamic Internet environment faces challenges.To address above issues,we introduce the public open ontology as the semantic support for describing the functional semantics of Web services,addressing the heterogeneity problem among different domain ontologies,and constructing executable solutions for given user requests.Achievements of this research are summarized as follows:(1)We present the augmentation of semantics of Web services based on public open ontology.More specifically,we analyze the distribution of implicit semantic information in Web services by applying natural language processing and statistical methods.And then,we enhance the semantics of Web services by associating inputs and outputs in Web services with concepts from public open ontology.Subsequently,we present two methodologies based on match degree and service discovery respectively to evaluate the quality of semantic annotation.(2)We consider the ontology matching by employing a knowledge base as the semantic bridge.More specifically,we formally define the ontology matching problem.And the we propose a four-stage ontology matching framework,including mapping terms in domain ontologies to equivalence and generalization terms in the knowledge base,matching terms from different domain ontologies on the basis of their corresponding terms in the knowledge base,direct matching as a supplementary mechanism in case that certain domain terms cannot be mapped to the knowledge base,and generating the closure of semantic correspondences by employing axioms in domain ontologies.(3)We study the interoperability among servibes based on location constraints.More specifically,we show the non-trivial presence of location constraints in Web Service by performing an empirical investigation.Based on social tags labeled on Web services,we propose an approach for automatically clarifying location constraints on Web services.And then we introduce the concept of expanded parameters,guided by the concept of service network,we propose service matching approaches based on location constraints with lower time complexity,which further guarantee the correctness of association relations among services.Finally,we construct initial composition based on association relations,generate the minimal composition alternatives based on the initial composition,and produce executable solutions for given user requests.In summary,we concentrate on issues of semantic descriptions of Web services and interoperability among Web services.Based on public open ontology,we study the functional semantics of Web services,overcome the conflict of semantic understanding among different service providers and requesters,and address the semantic sharing issues in service interoperability,thus lay the foundation for intelligent management,discovery and organization in the Internet environment.