Ontological Modeling Language Specification

Abstract

The Ontological Modeling Language, OML, is a thin vocabulary extension of W3C's OWL 2 Web Ontology Language designed to support precise and modular reasoning about descriptive systems engineering models. An OML Module maps to an OWL2-DL Ontology with SWRL rules both structurally and semantically. Structurally, an OWL2-DL Ontology provides classes, properties, individuals and data values that are represented in Semantic Web documents. These constructs form the basis of the OWL2-DL+SWRL mapping of an OML Module whose corresponding constructs include OML Entity, OML EntityRelationship, OML Datatype, OML DataProperty and OML TerminologyInstanceAssertion.

Vocabularies for Model-Based Systems Engineering (MBSE)

What is a vocabulary?

Quote from W3C's SemanticWeb summary:

On the Semantic Web, vocabularies define the concepts and relationships (also referred to as “terms”) used to describe and represent an area of concern. Vocabularies are used to classify the terms that can be used in a particular application, characterize possible relationships, and define possible constraints on using those terms. In practice, vocabularies can be very complex (with several thousands of terms) or very simple (describing one or two concepts only). There is no clear division between what is referred to as “vocabularies” and “ontologies”. The trend is to use the word “ontology” for more complex, and possibly quite formal collection of terms, whereas “vocabulary” is used when such strict formalism is not necessarily used or only in a very loose sense. Vocabularies are the basic building blocks for inference techniques on the Semantic Web.

The OML vocabulary is a subset of the OWL2-DL vocabulary for domain-specific MBSE

Defining a controlled vocabulary of a particular domain of systems engineering facilitates communication among stakeholders. In OML, a controlled vocabulary is an OML TerminologyBox, the semantics of which is defined by its mapping to a corresponding OWL2-DL Ontology. An OML TerminologyBox defines OML Terms and specifies constraints among them via OML Axioms. OML Terms and OML Axioms map to patterns involving [OWL2 Entities] and OWL2 Axioms.

Key properties of the semantics of OML in OWL2-DL+SWRL.

The semantics of an OML Module is defined by its mapping to a corresponding OWL2-DL Ontology with SWRL rules. OML is designed to provide support for modular and collaborative practices of domain-specific Model-Based Systems Engineering with an emphasis on precision and rigor thanks to the following key features of OML:

  • Mapping to OWL2-DL Ontologies {semantic-mapping}

    For collaborative modeling, every OML Module model that maps to a well-formed OWL2-DL Ontology with SWRL rules has a unique in the OML tabular interchange representation. Differences in the OML tabular interchange representation (e.g., with git diff) are isomorphic to differences in the structure of corresponding OML Modules.

  • Rigorous authoring of domain-specific vocabularies {rigorous-authoring}

    For authoring a rigorous domain-specific vocabulary in MBSE as an OML TerminologyBox, OML relies on reasoning with OWL2-DL with SWRL rules for verifying essential characteristics, including the satisfiability of every OML ConceptualEntity with respect to the OWL2 Axioms encoding the mapping of each OML Term and OML Axiom defined in the OML TerminologyBox.

  • Open vs. Closed World {open-vs-closed-world}

    For organizing rigorous domain-specific vocabularies in a modular, reusable fashion, OML distinguishes between two intended semantics of a particular OML TerminologyBox: open-world vs. closed-world. This distinction simplifies the problem of organizing OWL2-DL Ontologies with SWRL rules in a modular fashion such that reasoning over modular OWL2-DL Ontologies with SWRL rules can be correspondingly organized in a modular fashion where the inferences computed from sub-modules can be reused for sound OWL2-DL reasoning with SWRL rules over an aggregating super module.

  • A formalization of composite structure {composite-structure}

    For MBSE specifically, OML formalizes the notion of composite structure defined in the OMG UML 2.5, which is extensively used in the practice of MBSE with the OMG SysML 1.4 in so-called SysML Block Definition Diagrams (BDDs) that, in OML, correspond to an open-world OML TerminologyBox description of the possible structure of a system vs. SysML Internal Block Diagrams (IBDs) or SysML Parametric Diagrams (PARs) that, in OML, correspond to a closed-world OML TerminologyBox designation of a particular system structure topology. In SysML, IBDs and PARs reflect and implicit UNA that, in OML, is encoded via two distinct mechanisms: closed OML TerminologyBox where the semantics of each OML ConceptualEntity is restricted to be a singleton individual and OML DescriptionBox where a particular configuration of singleton individuals describe the topology of the structure of a particular system in a given state of affairs.

  • Modeling guidance, repair, verification and variation {guidance}

    For modeling guidance, repair, verification and variation analysis, OML is designed to facilitate using reasoning techniques OWL2-DL with SWRL rules for completing partial models (i.e., OML DesignationBox) where, for example, unsatisfiability precisely indicates missing essential information such as a value for an isIdentityCriteria=true or existentially restricted OML EntityDataProperty or a missing OML TerminologyInstanceAssertion for the domain, range or triple of an essential or existentially-restricted OML EntityRelationship.

  • Clear separation of provenance between assertions and entailments {separation-of-assertions-and-entailments}

    For facilitating the understanding of reasoning errors such as unsatisfiability and inconsistency, OML provides support for adopting a rigorous modeling methodology where human systems engineers have the unique responsibility for authoring essential descriptions that cannot be mechanically inferred and where automated reasoning have the unique responsibility for inferring derivable, non-essential descriptions without any possibility of inferring what should be human-authored essential descriptions. The key mechanism for this hinges on the fact that every OML ConceptualEntity (i.e., an OML Concept or an OML ReifiedRelationship) defined in the vocabulary of an OML TerminologyBox is, by definition, an essential definition. That is, the only mechanism to describe an instance of an OML ConceptualEntity is an OML ConceptualEntitySingletonInstance, which maps to an OWL2 NamedIndividual. Since OWL2-DL reasoners cannot create a fresh OWL2 NamedIndividual, it is therefore impossible for an OWL2-DL reasoner with SWRL rules to conjure a new OML ConceptualEntitySingletonInstance. On the other hand, an OML UnreifiedRelationship maps to an [OWL2 ObjectProperty], an instance of which is an OML UnreifiedRelationshipInstanceTuple that corresponds to a triple (see OWL2 RDF-based Semantic Conditions on Property Vocabulary) that any OWL2-DL reasoner with SWRL rules can infer as part of computing the entailments of an OWL2-DL Ontology. Thus, by partitioning the methodological use of OML ConceptualEntitySingletonInstance for representing human-authored system descriptions and that of OML UnreifiedRelationshipInstanceTuple for automated-reasoning entailments about such descriptions, a rigorous methodology can use OML to ensure clear provenance about each axiom implicated in a reasoner-provided explanation about an unsatisfiability or inconsistency result.

  • Change Management for collaborative modeling {change-management}

    For managing changes in the evolving descriptions and analyses of complex systems, the separation between essential descriptions (e.g., OML ConceptualEntitySingletonInstance) and derivable descriptions (e.g., OML UnreifiedRelationshipInstanceTuple) reduces the burden of change management to that of the essential descriptions and also provides a useful mechanism to assess the impact of changes to essential descriptions have on entailed non-essential descriptions. This impact analysis is particularly useful in conjunction with verification properties asserted using SWRL rules where the characteristics of an OML EntityRelationship such as transitivity, reflexivity and asymetry provide useful constraints for encoding characteristics of the domain-specific vocabulary or of that of a description of a particular system topology.

Organization of this Document

Since OWL2-DL is fundamentally important for the structure and semantics of OML, the organization of this document parallels that of the OWL2 Structural Specification.

results matching ""

    No results matching ""