Esm2005 Talk

8/13/2019 Esm2005 Talk

1/53

The need for Yasper

A closer look at Yasper

Integration

Conclusion

Yet Another Smart Process EditoR

Kees van Hee Reinier Post Lou Somers

Architecture of Information SystemsDepartment of Mathematics and Computer Science

TU Eindhoven

ESM 2005, Porto

Kees van Hee, Reinier Post, Lou Somers Yet Another Smart Process EditoR
http://www.tue.nl/http://www.tue.nl/http://find/

8/13/2019 Esm2005 Talk

2/53

The need for Yasper


Integration

Conclusion

Outline of this talk

1 The need for Yasper

Petri nets for process modelling

Why create Yasper?

2 A closer look at YasperYaspers modelling features

Simulation in Yasper

Implementation notes

3 IntegrationYasper and other tools

4 Conclusion

http://find/http://goback/

8/13/2019 Esm2005 Talk

3/53

The need for Yasper


Integration

Conclusion




Why create Yasper?





4 Conclusion

http://find/

8/13/2019 Esm2005 Talk

4/53

The need for Yasper


Integration

Conclusion




Why create Yasper?





4 Conclusion

http://find/

8/13/2019 Esm2005 Talk

5/53

The need for Yasper


Integration

Conclusion




Why create Yasper?





4 Conclusion


Th d f Y
http://find/

8/13/2019 Esm2005 Talk

6/53

The need for Yasper


Integration

Conclusion


Why create Yasper?




Why create Yasper?





4 Conclusion


Th d f Y
http://find/

8/13/2019 Esm2005 Talk

7/53

The need for Yasper


Integration

Conclusion


Why create Yasper?

Systems modelling

Our main interests:

business process modelling

software systems modelling

An adequate modelling technique is

clear

powerfulexact

well-supported


The need for Yasper
http://find/

8/13/2019 Esm2005 Talk

8/53

The need for Yasper


Integration

Conclusion


Why create Yasper?

Process modelling

Process modelling must express concurrency / cooperation.

Petri netsare adequate,

but need better support.

Most processes areworkflow nets:

with fixed start and end points.


The need for Yasper
http://find/

8/13/2019 Esm2005 Talk

9/53

The need for Yasper


Integration

Conclusion


Why create Yasper?

YasperWhy Yet Another Smart Process EditoR?

Many Petri net tools exist,

mostly in the academic world.

Our past contribution: theExSpecTcoloured Petri net tool.

Reasons to create another tool,Yasper:

make workflows easy to simulate

make Petri nets more palatable (for industry)

Microsoft integration via .NET (for industry)

integrate with other tools


The need for Yasper
http://www.exspect.com/http://www.exspect.com/http://find/

8/13/2019 Esm2005 Talk

10/53

The need for Yasper


Integration

Conclusion


Why create Yasper?

Use cases for Yasper

project: Deloitte Industry Prints:

"best/standard practices" business process modelsused byDeloitteconsultants

without a good modelling technique many ambiguities and plain errors

project: OGO 2.2:

business software modelling/prototyping project

part of computer science curriculumpreviously done withExSpecT

many other uses


The need for Yasper
http://www.deloitte.com/http://www.exspect.com/http://www.exspect.com/http://www.deloitte.com/http://find/

8/13/2019 Esm2005 Talk

11/53

The need for Yasper


Integration

Conclusion

Yaspers modelling features






Why create Yasper?





4 Conclusion


The need for YasperY d lli f t
http://find/

8/13/2019 Esm2005 Talk

12/53

The need for Yasper


Integration

Conclusion




Modelling features: overviewbasic Petri net elements

element notation description

place condition or resource

transition event or action

arc process flow

token (object in a) condition


The need for YasperY d lli f t
http://find/

8/13/2019 Esm2005 Talk

13/53

p


Integration

Conclusion




Modelling features: overviewgeneral purpose extensions


subnet spread across multiple pages

xor choice (split / join)

role (-) executer / resource

store data involved

inhibitor negative condition (no tokens)

reset clear condition(cleartokens)


The need for YasperYaspers modelling features
http://find/

8/13/2019 Esm2005 Talk

14/53

p


Integration

Conclusion

Yasper s modelling features



Modelling features: overviewextensions for automatic simulation


time processing time

cost (-) processing cost

case preserves workflow case

emitor generates workflow case

collector terminates workflow case


http://find/

8/13/2019 Esm2005 Talk

15/53


Integration

Conclusion




A basic Petri netplaces, transitions, arcs, tokens

Getting fuel at a petrol station


http://find/

8/13/2019 Esm2005 Talk

16/53


Integration

Conclusion




The net extended with Yasper featuresexhibiting most of them

Getting fuel, with more accuracy


The need for Yasper

A l l k YYaspers modelling features
http://find/

8/13/2019 Esm2005 Talk

17/53


Integration

Conclusion




Extension: subnetsspreading content over multiple pages

A subnet element


The need for Yasper

A l l k t YYaspers modelling features
http://find/

8/13/2019 Esm2005 Talk

18/53


Integration

Conclusion

p g



Extension: subnetsspreading content over multiple pages

The interface: references to places outside


The need for Yasper

A closer look at YasperYaspers modelling features
http://find/

8/13/2019 Esm2005 Talk

19/53


Integration

Conclusion

p g



Extension: choiceas known from flowcharts, UML activity diagrams

Choice elements indicate alternatives


The need for Yasper

http://find/

8/13/2019 Esm2005 Talk

20/53


Integration

Conclusion



Extension: choiceas known from flowcharts, UML activity diagrams

A basic Petri net equivalent


The need for YasperA closer look at Yasper

http://find/

8/13/2019 Esm2005 Talk

21/53


Integration

Conclusion



Extension: storesindicate (not simulate) data manipulation

Stores: data involved in transitions



http://find/

8/13/2019 Esm2005 Talk

22/53


Integration

Conclusion



Extension: rolesexecuters / resources

Transitions can be executed by roles




Si l i i Y
http://find/

8/13/2019 Esm2005 Talk

23/53


Integration

Conclusion



Extension: transition time and costfixed or stochastic

Time and cost assignments (for automatic simulation)




Si l ti i Y
http://find/

8/13/2019 Esm2005 Talk

24/53

p

Integration

Conclusion



Extension: workflow casescase sensitive places, emitors, collectors

Emitors and collectors mark start and end of workflow

http://find/

8/13/2019 Esm2005 Talk

25/53




8/13/2019 Esm2005 Talk

26/53

Integration

Conclusion



Extension: workflow casescase sensitive places, emitors, collectors

Transitions match cases on input places




http://find/

8/13/2019 Esm2005 Talk

27/53

Integration

Conclusion



Extension: special arc types

An example with more complex flow logic



I i


http://find/

8/13/2019 Esm2005 Talk

28/53

Integration

Conclusion



Manual simulation in Yasper

Playing the token game in the diagram



I t ti


http://find/

8/13/2019 Esm2005 Talk

29/53

Integration

Conclusion

S u at o aspe


Automatic simulation in Yasper

Starting an automatic run



Integration


http://find/

8/13/2019 Esm2005 Talk

30/53

Integration

Conclusion

p


Automatic simulation in Yasper

The report



Integration


http://find/

8/13/2019 Esm2005 Talk

31/53

Integration

ConclusionImplementation notes

Yasper makes simulation work

In Yasper, simulations arebased on exact execution semantics

very easy to set up and run

http://find/

8/13/2019 Esm2005 Talk

32/53


Integration



8/13/2019 Esm2005 Talk

33/53

Integration


Yasper makes simulation work

In Yasper, simulations arebased on exact execution semantics

very easy to set up and run



Integration



I l i
http://find/

8/13/2019 Esm2005 Talk

34/53

Integration


Automatic simulation demonstrates errorsand with alarming frequency, too

An error repaired



Integration



I l t ti t
http://find/

8/13/2019 Esm2005 Talk

35/53

g


Simulation makes Yasper work

In Yasper, simulations

immediately pinpoint most modelling errors

pinpoint deadlocks / bottlenecks in the process itself

can estimate overall throughput and efficiency



Integration



http://find/

8/13/2019 Esm2005 Talk

36/53

g









Integration



http://find/

8/13/2019 Esm2005 Talk

37/53









Integration



http://find/

8/13/2019 Esm2005 Talk

38/53







http://find/

8/13/2019 Esm2005 Talk

39/53


Integration

C l i




8/13/2019 Esm2005 Talk

40/53

Conclusionp e e tat o otes

ArchitectureYaspers library dependencies

PNML

PNMLUtilities

Drawing

PNMLIO

Yasper

Stresser

YasperLite

Pnml2Any

YasperWorkflowExecuter

System

System.Data

System.Drawing

System.XML

System.Windows.Forms

PNMLImaging

PNMLSimulationReporter

{EPNML 1.1: XSD / RNG} {PNML 1.3.2: RNG}



Integration

Conclusion



http://find/

8/13/2019 Esm2005 Talk

41/53

Conclusionp

Availability

Yasper is partly free:

Yasper program completely free to use

(but dont sue us when it breaks)

code not free

(but talk to us when you want it)

Get Yasper fromwww.yasper.org



Integration

Conclusion

Yasper and other tools
http://www.yasper.org/http://www.yasper.org/http://find/

8/13/2019 Esm2005 Talk

42/53

Conclusion




Why create Yasper?





4 Conclusion



Integration

Conclusion


8/13/2019 Esm2005 Talk

43/53

Conclusion

Development strategy: integration

Yaspers design philosophy:

do not compete with other toolstrying to duplicate their features

work with them instead

by translation / calling


8/13/2019 Esm2005 Talk

44/53


Integration

Conclusion


8/13/2019 Esm2005 Talk

45/53

Conclusion

Areas of integration

the use of data ("color"):

interface Yasper with a "data manager"

Petri net model checking / verification:

interface Yasper with analysis tools

non-Petri net modelling techniques:

supply translations from/to Yasper



Integration

Conclusion


8/13/2019 Esm2005 Talk

46/53

Areas of integration

the use of data ("color"):

interface Yasper with a "data manager"

Petri net model checking / verification:

interface Yasper with analysis tools

non-Petri net modelling techniques:

supply translations from/to Yasper



Integration

Conclusion


8/13/2019 Esm2005 Talk

47/53

Methods of integration

reuse of code libraries:

see architecture diagram above

common file formats

PNML, extended PNML



Integration

Conclusion


8/13/2019 Esm2005 Talk

48/53

Examples of integration

analysis tools, e.g. Woflan, INA

workflow engine (Yasper/InfoPath)simulation-only (with BPMN modeller)

process model translations (e.g. UML activity diagrams,

ProVision, BPMN, BPEL, ARIS, CRL)



Integration

Conclusion
http://find/

8/13/2019 Esm2005 Talk

49/53




Why create Yasper?

2




3

IntegrationYasper and other tools

4 Conclusion



Integration

Conclusion
http://find/

8/13/2019 Esm2005 Talk

50/53

Findings

Yasper simplifies Petri-net based modelling and simulation

Yaspers simulation is of great benefit in modelling

integrating tools is hard, but pays off



Integration

Conclusion
http://find/

8/13/2019 Esm2005 Talk

51/53

Ongoing/future work on Yasper itself

Provide more convenient editing:

larger nets

transformations

consistency checks

http://find/

8/13/2019 Esm2005 Talk

52/53


Integration

Conclusion

8/13/2019 Esm2005 Talk

53/53

Thank you

Maarten Leurs

lots of Yasper programmingapplying Yasper at Deloitte

Andries van Dijk

support at DeloitteOlivia Oanea, Ivo Raedts, Jan Martijn van der Werf, a.o.

using Yasper, making suggestions, bug reportswriting related softwarehelp with this presentation

Till Tantau

the LATEX beamer package

this audience

any feedback you have

http://find/

Date post:	04-Jun-2018
Category:	Documents
Upload:	mohsindalvi87
View:	216 times
Download:	0 times

Esm2005 Talk

Documents