Modeling and Assessment of Production Printing Workflows … · 2005-09-12 · Production Printing...
Transcript of Modeling and Assessment of Production Printing Workflows … · 2005-09-12 · Production Printing...
Modeling and Assessment ofProduction Printing Workflows Using
Petri NetsRajuRaju GottumukkalaGottumukkala
Computer Science DepartmentComputer Science DepartmentLouisiana Tech UniversityLouisiana Tech University
Dr. Tong SunDr. Tong SunPrincipal ScientistPrincipal Scientist
Adaptive & Smart Document System LabAdaptive & Smart Document System LabXerox Innovation GroupXerox Innovation Group
September 7, 2005September 7, 2005BPM Conference PresentationBPM Conference Presentation
Nancy, FranceNancy, France
©2004 Xerox Corporation. Xerox®, The Document Company® and the digital X® are trademarks of Xerox Corporation. All rights reserved.
Challenges & Motivations
– Major paradigm shift in the printing industry:• From a traditional human-intensive manufacturing process into more
automated digital workflows• Product centric => Solution centric
– In automated digital workflows,• Ever increasing number of heterogeneous devices and software
applications seamlessly interoperate with each other• High-speed, large volume, highly concurrent execution
– Business value:• Improve customer current production process efficiency by migrating into
digital workflows• Help customer make better-informed business decision• Foresee the performance implications to meet changing business needs
(what-if analysis)
Why Using Petri Nets?
– Single FORMAL model for the process analysis,verification, and simulation
– Open source tools available (Renew, Design/CPN, etc.)
– Promising new development fronts• PNML
• Petri Net Ontology
• BPEL2PN
Production Printing Workflows: Core Domain Model
– Service Model• Built upon the SOA and Semantic Web Service notions
• Production Printing Service Specifics:
– Large grain logical model (functionality externally visible)
– Building block of a production printing workflow
– Document Model• Describes the entire lifecycle of the content data model
• Transfer modes: spool, stream.
– Job Model• Intent-based (what) vs. instruction-based (how)
• Print job dynamic operational behavior and lifecycle
Production Printing Workflows in Petri Nets(1)
– Executable Service Net (ESN)
Production Printing Workflows in Petri Nets(2)
– Executable Workflow Net (EWN)• A superset of ESNs = ESNs + Control Elements• Input to Renew tool to simulate process model
– a user given set of simulation input profiles: workload, jobarrival rate, document types, etc.
– service capability specification» Setup time» Processing time» Failure probability» Repair Time
– generate the process performance metrics:» Turnaround time & throughput» Waiting time» Service utilization rate
Production Printing Workflows in Petri Nets(3)
– Abstract Service Net (ASN)• A finite set of service states (as Places) and service
transitions (as Transitions)• Job is a Token• Document is characterized as Token properties
– Abstract Workflow Net (AWN)• A super net of ASNs = ASNs + Control Elements• Input to Design/CPN Tool for workflow model verification
(deadlock, liveness, reachability, etc.)
Simulation-driven Workflow Assessment Tool
3rdParty WorkflowModelingTool (s)
Petri Net Simulator(Renew)
Data Analysis Component
Simulation Components
GUI EditorSimulation resultreporter
EWN Generator
ESN Templates
Event Driven Simulation Engine
Workflow Simulation Interface
Workflow Model
Constructor
Database
Trace Generator
Scenario – Sample Process Models
Scenario – Sample Models (2)
Abstract Workflow Net
Executable Workflow Net
On-going & Future Work
– Seamless mapping/integration between standard XML-basedworkflow specifications (e.g. BPMN, BPEL, JDF) into Petri Net(or PNML)
– Extension scheme for Service Net (generic vs. specific)
– Extend the construction scheme of Abstract/ExecutableService Net based on Job lifecycle and Document lifecycle
– Extend the construction scheme of Abstract/ExecutableWorkflow Net based on Document’s semantic behavior
– Semantic Web, Web Service, OWL-S