A Semantic Based Workflow Management in a Virtual Organization

1
A Semantic Based Workflow Management in a Virtual
Organization

• Yun-Heh (Jessica) Chen-Burger
• AIAI, CISA, The University of Edinburgh
• e-Science Workflow Workshop
• Dec 3-5, 2003
• AIAI Commercial, AKT, CoAKTinG project

2
Work Areas

• Visual and declarative BPM language RACD, FBPML
• Automatic modelling support
• Model creation and documentation
• Ontology based knowledge sharing
• Analysis, critiquing and state-stepping
• Workflow system
• Guided modelling activities
• Guided business operation
• Visualisation aids
• Process view, product view, agent status view
• Process dependency view, information life cycle
  view
• Provides support for virtual organisations

3
Knowledge Based Support Framework for Workflow
Workflow System I
Model Translation
Model Reuse
Agent J1
Workflow System J
Formal Representation
Agent J2
GUI
Models
Modellers
End User
Broker
State Stepping
Verification, Validation Critiquing Within a
model And between models and ontology
Agent K1
Semantic Web Language Publishing
Workflow System K
Agent K2
End User
Distributed Environment
KBST-EM
4
AKT Research Map and Ontology

• OWL Representation
• http//www.aiai.ed.ac.uk/jessicac/project/akt/akt
  -map-owl.xml
• http//www.aiai.ed.ac.uk/jessicac/project/akt/akt
  -map-onto.xml

5
The Three-Layered Business Process Modelling
Approach
Organisation/ Business/ Goal Model
Application Layer
Goals and Policies
Logical Layer
Operational Requirements
Process/Data Model, Ontology
Library of Modules
System Layer
System Requirements
6
FBPML Notation
Screen Shot of KBST-EM
7
FBPML Process Ontology Skeleton

• Nodes
• Primitive Activity
• Activity
• Role
• Time Point
• Junctions
• And-junction
• Or-junction (inc. Xor-junction)
• Links
• Precedence-Link
• Synchronisation-Bar
• Action/Process Decomposition
• Task decomposition
• Alternative decomposition

8
Action Types (selective)

• Data manipulation actions
• Create(InstanceAttributeRelation)
• Delete(InstanceAttributeRelation)
• Update(InstanceAttributeRelation)
• Conditional actions
• Condition_action(Conditional_statement_list,
  Action_list)
• User interfaces
• Report(Title, Content)
• Read_user_input(Title, Input)
• Control of system operation (execution route)
• Construct issues (INCA ontology based)
• Communication with external agents/brokers
• Post_issue(Header, Issue_content),
  Receive_issue(Header, Issue_content)
• Update_event_status(Event, Status)
• Domain Functions
• Update_cost_in_event(Event, Cost)
• Update_solution_in_issue(Issue, Solution)
• Update_solution_in_solution(Solution, Solution)
• Store_customer_requirements(Requirement_content)

9
Additional Primitives

• Time and its manipulation
• Trigger
• Event
• Conditional statements
• Preconditions
• Postconditions
• Life cycle
• Event life status and cycle
• Process life status and cycle
• Each node has attributes, e.g.
• Process Instance_Id, Process_type, Life_status,
  Priority, Average_time_cost, Begin/End_time,
  Service_Requester/Service_Provider, Trigger,
  Preconditions, Actions, Postconditions.

10
Process Model at the Sales and Marketing Site
Screen shot from KBST-EM
11
Process Model at the Technical Site
Screen shot from KBST-EM
12
Notation for Ontology
13
Domain Ontology
Extended based on AKT Ontology, Compatible with
INCA Ontology
14
Abstraction of Collaboration ArchitectureIn the
PC Configuration Domain
Edinburgh Costing Site
Aberdeen Tech. Site
I-X Process Panel
I-X Process Panel
Com I-1
Com I-2
BPM - 1
KRAFT Constraint Solver
Com P-1
Workflow
Concept Mapping
INCA-FBPM Ontology
Constraint Ontology
Mapped Ontology
Partial Domain Ontology (tech)
Partial Domain Ontology (SM)
15
Mapped to and Operated using IX Process Panel
Web page and movie of live record
http//www.aktors.org/technologies/kraft-ix/
16
Example
start_junction( Confirm Entering Clearing
Process). process( Confirm Entering Clearing
Process, Instance, ccs(X), or(past_due_date(St
udent), not_successful(Student)), update_status(S
tudent, clearing) ).
link(Confirm Entering Clearing Process,
Execute Clearing Process).
17
Example Process

• process(Instance, 'Receive Customer Request'/1,
  created, Priority,
• _Begin_time/_End_time, 1,
• (Requester/Requester_type,
  edinburgh/pc_specification),
• event_occ(Instance, customer_request_for
  _pc_specification,
• received/_, Priority,
  _Begin/_End,
• (Requester/Requester_type,
  _Provider/_Provider_type),
• _Event_content) ,
• true,
• cond_action(not_exist(instance_of(Reques
  ter, customer)),
• create(instance_of(Requester
  , customer)) ),
• create(instance_att(Requester, event,
  Instance)),
• create(instance_of(Instance, event))
• ,
• exist(instance_of(Requester, customer)),
• exist(instance_att(Requester, event,
  Instance)),
• exist(instance_of(Instance, event))
• ).

18
Internal View of WFE
The User
External Event And Interactions
IDEF Methods
FBPML Process Model
RAD
PSL/PIF
WSFL
Process Agenda
Process Execution
BSDM-BM
Think Workflow Engine
Standardised Methods
FBPML-DL
Process, Event And Status
System State
data
Persistent Information Storage
World State Awareness Of Workflow Engine
19
Characteristics of FBPML

• Contains a process and data modelling languages,
  including ontology
• Provides an abstraction that is separated from
  the actual implementation
• Has precise execution semantics (that is grounded
  in data semantics) that supports generations of a
  workflow virtual machine at run time
• Provides a visual presentation of the process
  model
• Provides a visual presentation for the underlying
  Ontology
• visual data modelling languages may be ER, UML
  Data Diagram, etc)
• Has a notion of time that may be synchronised
• Suitable for a distributed environment
• Can link to organisation/business/goal model
• Provides a suitable foundation for automatic VV
  (static, simulation, within one model, between
  models), model critiquing, inferencing (e.g.
  dependencies), confirming with ontology,
  planning, scheduling, etc.

20
Three layer grid abstraction
Data to Knowledge
Control
Source Richard Kenway
21
Relating workflow to the three layered grid
abstraction
Semantic Description
Declarative Process Model
Virtual Workflow Machine Information/Message Pass
ing, Process Execution
Data to Knowledge
Realisation of Workflow
Data Transaction
Transaction of Workflow
adapted from Richard Kenway
22
Workflow linking to SemanticWeb Technologies

• FBPML-DL is mapped to OWL, including
• Ontology
• Domain model (data mode and instances)
• FBPML is currently mapped to BPML/WSCI and DAML-S
• To develop semantic grid, wisdom may be gained
  and lessons learned from experiences made from
  semantic web development
• Interoperability
• Loose coupling of heterogeneous systems that are
  geographically disperse
• Collaboration to achieve common and individual
  goals

23
Semantic Web Layer Cake
24
adapted from WS-CDL
25
Virtual Organisation with Workflow Systems
Intelligent Agent i
Constraint Solver w
Intelligent Agent - z
Internal Broker Y
Problem Solver k
Internal Broker j
Workflow System - 2
Workflow System - 1
SWL Wrapper
SWL Wrapper
Independent Broker Y
Cost Checker
Database m
Database
Other Broker/ Agents
Problem Solver x
(Sub-)Organisation A
(Sub-)Organisation B
26
Challenges and Issues for Standard Workflow
Languages

• Many emerging WF standard languages are competing
  
• Rapid evolution on semantic web languages
• Many layers of semantic web languages
• Standard WF languages often lack visual
  representation
• Lack of semantic support for workflow operations
• Standard services ontology ?
• Standard process ontology ?
• Lack of explicit representation and organisation
  of data manipulated by processes
• Lack of explicit representation of time and
  synchronisation of processes in addition to the
  beginning and ending of a process
• Lack of explicit support to operating context
• Lack of explicit representation and therefore
  transparent control of system environment that
  are being altered by processes
• Lack of linking to organisation/role/capability
  model
• Lack of explicit linking to goal/business model
• Lack of guidelines for producing good process
  model and rich modelling examples

27
End of TalkThank you for listening

• Yun-Heh Chen-Burger
• AIAI, CISA, The University of Edinburgh
• Workflow Handbook 2003, WfMC.
• Home page http//www.aiai.ed.ac.uk/jessicac
• PM http//www.aiai.ed.ac.uk/jessicac/project/pc-
  configuration-model/top-level.html
• WF http//www.aktors.org/technologies/workflow/
• WF http//www.aktors.org/technologies/kraft-ix/