© Geodise Project, University of Southampton, 2002.

Knowledge Managementin Geodise

Geodise Knowledge Management Team

Barry Tao, Colin Puleston, Liming Chen, Paul Smart

University of SouthamptonUniversity of Manchester

Epistemics Ltd.

© Geodise Project, University of Southampton, 2002.

Distinguishing features of the architecture:Tackling the six challenges of knowledge management life cycle in an integrated framework.A layered modular structure with each component dealing with a specific aspect of knowledge engineering process.Supporting the exploitation of different techniques and tools. Each of them can be updated while others kept intact. Using ontologies to generate machine-interpretable semantically-enriched content and also knowledge bases.A knowledge portal provides mechanisms for knowledge maintenance and resource control.

Easy knowledge reuse and sharing through web service technology over the Internet.Flexible - adding knowledge any time through portal.

Service-oriented approach - separating domain knowledge from operational knowledge.

Extensible - plugging in new functions as knowledge services.Robust - adopting new techniques/tools while keeping the system functioning.

While computing increasingly addresses collaboration, sharing and interaction with the powerful support of the emerging distributed computing infrastructures such as web services and Grid technologies, there is a growing demand for ontology and knowledge technologies to provide semantic support for service integration, the sharing and coordinated use of knowledge across distributed, heterogeneous, dynamic virtual organisations. We have developed and partially implemented an architecture to provide knowledge services. It has been applied to Grid-enabled design search and optimisation (Geodise).

Knowledge Management Architecture for Grid Computing

© Geodise Project, University of Southampton, 2002.

Knowledge acquisition and modelling lay the foundation for the knowledge service architecture, which produces ontologies and the types of template and structure where knowledge can be held. It has been carried out by the CommonKADS knowledge engineering methodology and the PC PACK knowledge management integrated tools.

Concept LadderProtocol Editor

Diagram Tool

The aim of knowledge portal is:To make knowledge available and accessibleTo provide tools for knowledge reuse and exchangeTo provide security infrastructureTo manage knowledge resourceTo support online forum and maintain mailing listsTo disseminate the latest advance of the domain

Current functions of knowledge portal include:Security mechanismKnowledge publicationService registrationKnowledge retrieval and updateService application informationResource management via version control

Security Control

Knowledge PortalBrowser-based Interface

© Geodise Project, University of Southampton, 2002.

Knowledge representation and publishing is to represent knowledge in a well-structured, well-indexed form and make them ready for sharing and use. Geodise domain knowledge has been represented in CommonKADS task models, concept hierarchy and a design workflow and published in the form of knowledge web and XML format.

DesignWorkflow

Knowledge Web

Knowledge Web in XML

Geodise Domain Models

Ontologies serve as the conceptual backbone for knowledge sharing and management in the above integrated architecture for knowledge services.We have developed ontologies for design search and optimisation using both Protégé and OilEd ontology tools.

Protégé Ontology Concept Instances

Protégé Ontologyin HTML Format

Protégé OntologyConcept Hierarchy

OilEd Ontology in DAMLClasses, Hierarchy & Properties

Geodise Knowledge Models

© Geodise Project, University of Southampton, 2002.

Ontology service provides a Java API giving full access to any DAML+OIL ontology that is available over the Internet. The API is exposed through both a SOAP-based web service and a CGI interface for common ontological operations, such as subsumption checking, navigating concept hierarchy and retrieving concept and attributes.

Ontology Service CGI Interface

SOAP-based Ontology Service WSDL

Ontology ServicePresentation in Portal

Annotation is necessary to add semantic content to document or website, thus facilitating information sharing, reuse and automatic machine processing. We have used OntoMat-Annotizer annotation tool in Geodise to annotate the workflow for particular design problems and then save them in a knowledge base. The semantically enriched archive can then be queried, indexed and reused later to guide future designs.

Annotation Process

Original Document

Metadata Added after Annotation

Ontology Services &Semantic Annotations

© Geodise Project, University of Southampton, 2002.

Knowledge-based Ontology-assisted Workflow Construction ArchitectureWorkflow construction

Use ontology & ontology services to facilitate method selection & instantiation

Design advice Use process knowledge

& advice reasoner to provide design advice

Storage & query Add semantics to

databases & workflow automatically, thus facilitate content-based search and query

© Geodise Project, University of Southampton, 2002.

Ontology service

Ontology-driven Workflow Construction Ontology-assisted task navigation & selection (Drag and Drop) Ontology assisted task configuration

Task ontology

© Geodise Project, University of Southampton, 2002.

Knowledge-based Systems for EDSO

Gambit journal

file editor

Knowledge-based advisor

Design advice

Add a task

Process-level design advisor Service-oriented paradigm Ontology as common terms

Task-level design tools Ontology-assisted Gambit journal file

editor Critique on commands & workflow

Knowledge APIs XML-based messaging

© Geodise Project, University of Southampton, 2002.

% Query and locate the instance fileresult=gd_query('standard.archiveDate > 2003-03-16');ProblemID=result{1}.standard.ID;local_file_path=gd_retrieve(ProblemID,'d:\'); %local_file_path=’D:/geodise/XML_Templates/problem_mo_arcadia5.xml’

% Specify the local path of the problem profile instance.problem_profile_instance=local_file_path;

% get information about design variablesxp=knowledgeapi.XMLParser(problem_profile_instance);% get information about design variablesdvs=knowledgeapi.DesignVariables1(xp.getDoc);%get information about objective functionof=knowledgeapi.ObjectiveFunction(xp.getDoc);% The recommented boundaries for the design parameters, useful as% allows the user to use a constrained optimisation.% design parameter boundsdsgnmin = rot90(dvs.getLowerBounds);%[ -0.15 0.40 2.00 0.00 0.50 2.00];dsgnmax = rot90(dvs.getUpperBounds);%[ 0.05 0.80 10.00 0.15 0.85 5.00];defaultValues=rot90(dvs.getDefaultValues);% design parameters selected to be design variablesselect = rot90(dvs.getSelected);%[3,5];selectedObjName=char(of.getSelectedObjName);

% Create a setup file for the optimisation% [setup_struct,setupFileID] = arcadia5_setup( [3,5],'peakvel2','',[0.04,0.5,3,0.02,0.6,3],1.4,1.5,0.1,4.6,[ -0.75,1.5,0,0.95],[0.05,0.0125,0.05,0.0125])[setup_struct,setupFileID] = arcadia5_setup( select,selectedObjName,'',defaultValues,1.4,1.5,0.1,4.6,[ -0.75,1.5,0,0.95],[0.05,0.0125,0.05,0.0125])

% Example use of the programDesignVariables= rot90(dvs.getSelectedDefaults) %[2.0, 0.85];

XML-Based Component Management XML & XML Schema

Java/JAXFront technology

Access via knowledge APIs

Potential ontology support Visualisati

on Implementation

Framework

<ProblemProfile description="Arcadia5 design problem" dg_id="" lastTimeUsed="2003-03-04T11:21:36" timeCreated="2002-11-23T09:20:23"user="barry" xmlns="http://www.geodise.org/knowledge" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.geodise.org/knowledgeD:\geodise\XML_Templates\problem_mo.xsd">

<designVariables><name>a_l</name><meaning>The maximum bump height of the Hicks-Henne bump function on the lower surface of the nacelle</meaning><unit>mm</unit><limit>

<continousLimit><lower_bound>-0.15</lower_bound><default_value>0.05</default_value><upper_bound>0.04</upper_bound>

</continousLimit></limit><fixed>true</fixed>

</designVariables><designVariables>

<name>xp_l</name><meaning>The bump peak location (on the x-axis) of the Hicks-Henne bump function on the lower surface of the nacelle</meaning><unit>mm</unit><limit>

<continousLimit><lower_bound>0.40</lower_bound><default_value>0.5</default_value><upper_bound>0.80</upper_bound>

</continousLimit></limit><fixed>true</fixed>

</designVariables><designVariables>

<name>t_l</name><meaning>The bump width parameter of the Hicks-Henne bump function on the lower surface of the nacelle</meaning><unit>mm</unit><limit>

<continousLimit><lower_bound>2.00</lower_bound><default_value>3</default_value><upper_bound>10.00</upper_bound>

</continousLimit></limit><fixed>false</fixed>

</designVariables><designVariables>

<name>a_u</name><meaning>The maximum bump height of the Hicks-Henne bump function on the upper surface of the nacelle</meaning><unit>mm</unit><limit>

<continousLimit><lower_bound>0.00</lower_bound>

Problem

Definitio

n

Design

Variable

DefinitionProblem Instance in

XML

Use Problem Instances

in Matlab

Framework

© Geodise Project, University of Southampton, 2002.

Semantic Component Management

PSE:Workflow Construction

Functions in .m files Matlab ExecutionEnvironment

Semantic DescriptionUsing DAML-S

Enactment or Mapping to .m Scripts

Ontology Services

Resource annotations

Geodise ServiceOntology

Geodise Domain Ontologies

Knowledge RepositoriesWith Embedded Semantics

(RDF Triple Store)

Inference engines

Retrieval andQuery APIs

To knowledge-enable

EDSO Ontologies (service/function)

Ontology Services

Service/FunctionForm or Templates

Semantics-based Query & Inference

Semantics-based Query & Inference

RDF Triple Store& Permanent Storage (DBS)

Concept Java Classes

RDF GeneratorJena

RD

F A

PIs

Geodise Users

Create

Re-use

Geodise toolkit in Matlab

• Common interface• Semantic content• Machine understandable• Flexible discovery & composition

Service-oriented semantic descriptions

Automated form generation

RDF as the representation formalism

RDF triple store & DBMS

Semantic & knowledge technologies

Implementatio

n framework

© Geodise Project, University of Southampton, 2002.

Ontology Delivery

Classification

Language

Indexing

FaCT

Terminology Service

OilEd

User Interfacetoolkits

Browser & viewers,Query builders,Form builders,Wizards

Annotation & Linking(COHSE)

Instance Store

ViewsMatcher

Core OntologyService (OS)

© Geodise Project, University of Southampton, 2002.

Ontology Building Environments

Knowledge acquisitionOntology creation and maintenanceOntology parsers and checkersOntology alignment and mergingOntology reasoning

© Geodise Project, University of Southampton, 2002.

Ontology View FrameworkOtherViews

Other Views??

Ontology Client

Ontology Server

WEB

Semantic Network View (Configurable)

DAML+OIL/OWL Ontology

Instance Store (Database)

GeodiseTasks

Geodise Concepts

FaCTReasoner

GONGConcepts

Concept Query View

© Geodise Project, University of Southampton, 2002.

Geodise Task Ontology View

Ontology Client

API

GUI

Geodise Task View

GeodiseWorkflow Application

Semantic Network View

Semantic NetConfiguration

Semantic NetAccess

Configured for Geodise task ontology

Configured for Geodise task

ontology