1. Modelling & Mining Event-based Concurrent
Declarative Processes as
Dynamic Condition Response (DCR) Graphs
Thomas T. Hildebrandt
joint work with
S. Debois, K. R. Ulrik, P. H. Laursen (ITU),T. Slaats (KU), R. Mukkamala (CBS), M. Marquard (Exformatics)
Process Intelligence, Modelling & Optimisation Group
IT University of Copenhagen (ITU) Denmark
Invited talk - ATAED 2017 @ PN/ACSD2017
June26th, 2017
IT UNIVERSITY OF COPENHAGENITUNIVERSITYOFCOPENHAGEN SUBMISSION OF WRITTEN WORK
Class code:
Name of course:
Danish Defence Agency
2. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Motivation
2
IT systems increasingly control and
support critical processes & interactions
between humans and machines
3. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Motivation
2
IT systems increasingly control and
support critical processes & interactions
between humans and machines
and operate in
unpredictable and
changing contexts
4. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Motivation
2
IT systems increasingly control and
support critical processes & interactions
between humans and machines
Need for flexibility, adaptability & compliance with
legal, safety and security regulations
and operate in
unpredictable and
changing contexts
5. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
40 years of flow diagrams…
3
zc
:=
Computer Science and
Office Information Systems
By Clarence A. Ellis and Gary J. Nutt
ORDERPROCESSING
LogRequest
TypeOrder
SendOrder
ReceiveOrder
BrocessII
J
I
I
•
:1.1
Customer
Request
Arrival
lL
-""/
I
I
r
•r
J/.I
I
A/I
"
Order
Form
'"I
+0I,Custamer
JjFile
I
IIBillingFile
I
I
I
J
I
J
I
I
I
I
I
I
I
/
I,
lOut"
1:;;1tJ.lOut.I
tForm'--------/--""
',----..._------_.._-..,.-'
F.igure2
70’ties:
2010 - now:
6. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
What is the problem with flows?
4
7. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
What is the problem with flows?
4
Inflexible execution and development
IT UN
Towards effective, flexible & legally compliant digital knowledge workflows
Thomas T. Hildebrandt (hilde@itu.dk)
The computer says no….
5
Performance
goals
Best
practice
the computer
says no
The baby is
coming!
8. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
What is the problem with flows?
4
Only describe how not why
difficult to adapt when regulations change
Inflexible execution and development
IT UN
Towards effective, flexible & legally compliant digital knowledge workflows
Thomas T. Hildebrandt (hilde@itu.dk)
The computer says no….
5
Performance
goals
Best
practice
the computer
says no
The baby is
coming!
9. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
What is the problem with flows?
4
Only describe how not why
difficult to adapt when regulations change
Limited support for modularity & re-use
Inflexible execution and development
IT UN
Towards effective, flexible & legally compliant digital knowledge workflows
Thomas T. Hildebrandt (hilde@itu.dk)
The computer says no….
5
Performance
goals
Best
practice
the computer
says no
The baby is
coming!
10. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
The true process landscape
5
Completely predictable &
many repetitions, few changes
Completely unpredictable,
few repetitions & changes
Partially predictable,
some repetition
& some changes
Very few (<5%) work processes fully predictable
Workflows and it-systems contain unpredictable parts and
need to be changed over time
A FUTURE
THAT WORKS:
the impact of automation
in Denmark
11. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Flexibility versus Support
6
Motivation
Flexibility versus Support in workflows
• Flexibility: ability to defer, change,
and deviate
• support: provide analysis and
guidance
• unstructured: do what ever you
want, but get no support
• structured: support, but no
flexibilityClassical trade-off between flexibility and support1
[1] W.M.P. van der Aalst et al. Declarative workflows: Balancing between flexibility and support
Sunday, March 14, 2010
Motivation
Flexibility versus Support in workflows
• Flexibility: ability to defer, change,
and deviate
• support: provide analysis and
guidance
• unstructured: do what ever you
want, but get no support
• structured: support, but no
flexibilityClassical trade-off between flexibility and support1
[1]W.M.P. van der Aalst et al. Declarative workflows: Balancing between flexibility and support
unday, March 14, 2010
[Schmidt & Bannon:Taking CSCW Seriously: Supporting Articulation Work, 1992]
Already in 1983, researchers in
Computer Supported Cooperative
Work (CSCW) concluded that office
automation systems “do not deal
well with unanticipated
conditions” (Barber) & “were
automating a fiction” (Sheil)
12. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Flexibility versus Support
6
Motivation
Flexibility versus Support in workflows
• Flexibility: ability to defer, change,
and deviate
• support: provide analysis and
guidance
• unstructured: do what ever you
want, but get no support
• structured: support, but no
flexibilityClassical trade-off between flexibility and support1
[1] W.M.P. van der Aalst et al. Declarative workflows: Balancing between flexibility and support
Sunday, March 14, 2010
Motivation
Flexibility versus Support in workflows
• Flexibility: ability to defer, change,
and deviate
• support: provide analysis and
guidance
• unstructured: do what ever you
want, but get no support
• structured: support, but no
flexibilityClassical trade-off between flexibility and support1
[1]W.M.P. van der Aalst et al. Declarative workflows: Balancing between flexibility and support
unday, March 14, 2010
[Schmidt & Bannon:Taking CSCW Seriously: Supporting Articulation Work, 1992]
Already in 1983, researchers in
Computer Supported Cooperative
Work (CSCW) concluded that office
automation systems “do not deal
well with unanticipated
conditions” (Barber) & “were
automating a fiction” (Sheil)
“Good standards for business process
modelling are still missing and even today’s
WFMSs are too rigid”
Process-Aware Information Systems:
Design, Enactment, and Analysis
Wil M.P. van der Aalst
Department of Mathematics and Computer Science, Eindhoven University o
nology, P.O. Box 513, NL-5600 MB Eindhoven, w.m.p.v.d.aalst@tue.nl
Abstract. Process-aware information systems support operational busine
cesses by combining advances in information technology with recent in
13. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Example from local government
7
Compliant?
• Lov om Aktiv beskæftigelsesindsats
(LBK nr 1428 af 14/12/2009)
• Lov om Aktiv socialpolitik
(LBK nr 946 af 01/10/2009)
• Lov om Arbejdsløshedsforsikring
(LBK nr 574 af 27/05/2010)
• Lov om Integration af udlændinge
(LBK nr 1062 af 20/08/2010)
• Lov om Sygedagpenge
(LOV nr 563 af 09/06/2006)
• Retssikkerhedsloven
(LBK nr 1054 af 07/09/2010)
• Datagrundlag
(BEK nr 418 af 23/04/2010)
arbejdsgangsbanken.dk2010: Case Studies of Best
Practice Workflow and
Workflow in Practice
(Innovation Network Project)
14. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Example from local government
7
Change in law Compliant?
• Lov om Aktiv beskæftigelsesindsats
(LBK nr 1428 af 14/12/2009)
• Lov om Aktiv socialpolitik
(LBK nr 946 af 01/10/2009)
• Lov om Arbejdsløshedsforsikring
(LBK nr 574 af 27/05/2010)
• Lov om Integration af udlændinge
(LBK nr 1062 af 20/08/2010)
• Lov om Sygedagpenge
(LOV nr 563 af 09/06/2006)
• Retssikkerhedsloven
(LBK nr 1054 af 07/09/2010)
• Datagrundlag
(BEK nr 418 af 23/04/2010)
arbejdsgangsbanken.dk2010: Case Studies of Best
Practice Workflow and
Workflow in Practice
(Innovation Network Project)
15. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Example from local government
7
Change in law Compliant?
• Lov om Aktiv beskæftigelsesindsats
(LBK nr 1428 af 14/12/2009)
• Lov om Aktiv socialpolitik
(LBK nr 946 af 01/10/2009)
• Lov om Arbejdsløshedsforsikring
(LBK nr 574 af 27/05/2010)
• Lov om Integration af udlændinge
(LBK nr 1062 af 20/08/2010)
• Lov om Sygedagpenge
(LOV nr 563 af 09/06/2006)
• Retssikkerhedsloven
(LBK nr 1054 af 07/09/2010)
• Datagrundlag
(BEK nr 418 af 23/04/2010)
How to change??
arbejdsgangsbanken.dk2010: Case Studies of Best
Practice Workflow and
Workflow in Practice
(Innovation Network Project)
16. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Example from local government
7
Change in law Compliant?
• Lov om Aktiv beskæftigelsesindsats
(LBK nr 1428 af 14/12/2009)
• Lov om Aktiv socialpolitik
(LBK nr 946 af 01/10/2009)
• Lov om Arbejdsløshedsforsikring
(LBK nr 574 af 27/05/2010)
• Lov om Integration af udlændinge
(LBK nr 1062 af 20/08/2010)
• Lov om Sygedagpenge
(LOV nr 563 af 09/06/2006)
• Retssikkerhedsloven
(LBK nr 1054 af 07/09/2010)
• Datagrundlag
(BEK nr 418 af 23/04/2010)
How to change??
arbejdsgangsbanken.dk
Change in practice
2010: Case Studies of Best
Practice Workflow and
Workflow in Practice
(Innovation Network Project)
17. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Example from local government
7
Change in law Compliant?
• Lov om Aktiv beskæftigelsesindsats
(LBK nr 1428 af 14/12/2009)
• Lov om Aktiv socialpolitik
(LBK nr 946 af 01/10/2009)
• Lov om Arbejdsløshedsforsikring
(LBK nr 574 af 27/05/2010)
• Lov om Integration af udlændinge
(LBK nr 1062 af 20/08/2010)
• Lov om Sygedagpenge
(LOV nr 563 af 09/06/2006)
• Retssikkerhedsloven
(LBK nr 1054 af 07/09/2010)
• Datagrundlag
(BEK nr 418 af 23/04/2010)
How to change??
arbejdsgangsbanken.dk
Change in practice
Still compliant ??
2010: Case Studies of Best
Practice Workflow and
Workflow in Practice
(Innovation Network Project)
18. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Example from local government
7
Change in law Compliant?
• Lov om Aktiv beskæftigelsesindsats
(LBK nr 1428 af 14/12/2009)
• Lov om Aktiv socialpolitik
(LBK nr 946 af 01/10/2009)
• Lov om Arbejdsløshedsforsikring
(LBK nr 574 af 27/05/2010)
• Lov om Integration af udlændinge
(LBK nr 1062 af 20/08/2010)
• Lov om Sygedagpenge
(LOV nr 563 af 09/06/2006)
• Retssikkerhedsloven
(LBK nr 1054 af 07/09/2010)
• Datagrundlag
(BEK nr 418 af 23/04/2010)
How to change??
arbejdsgangsbanken.dk
Change in practice
Still compliant ??
2010: Case Studies of Best
Practice Workflow and
Workflow in Practice
(Innovation Network Project)
gave up January 2013
19. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Flow diagrams are like…
8
20. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Flow diagrams are like…
8
following pre-specified route(s)
21. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Flow diagrams are like…
8
following pre-specified route(s)
If you leave the route,
you are on your own &
can not see the road
22. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Flow diagrams are like…
8
following pre-specified route(s)
If you leave the route,
you are on your own &
can not see the road
If the map changes,
you have no idea how to
update the routes
23. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
We need a workflow GPS
9
24. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
We need a workflow GPS
9
where the route is calculated
from the map and your goal
25. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
We need a workflow GPS
9
where the route is calculated
from the map and your goal
If you leave the route,
a new one can be calculated
26. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
We need a workflow GPS
9
where the route is calculated
from the map and your goal
If you leave the route,
a new one can be calculated
If the map changes,
the route can be adjusted
27. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
We need a workflow GPS
9
where the route is calculated
from the map and your goal
If you leave the route,
a new one can be calculated
If the map changes,
the route can be adjusted
- achieved by modelling processes
declaratively as DCR Graphs
28. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
DCR Graphs: A decade of R&D
10
2007-11: Computer Supported Mobile Adaptive Business Processes
(Danish Research Foundation for Technology and Production)
29. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
DCR Graphs: A decade of R&D
10
2007-11: Computer Supported Mobile Adaptive Business Processes
(Danish Research Foundation for Technology and Production)
2008-2012: Trustworthy Pervasive Healthcare Processes
(Danish Strategic Research Council)
30. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
DCR Graphs: A decade of R&D
10
2007-11: Computer Supported Mobile Adaptive Business Processes
(Danish Research Foundation for Technology and Production)
2008-2012: Trustworthy Pervasive Healthcare Processes
(Danish Strategic Research Council)
2009-13: Services in Context
(Danish Strategic Research Council)
31. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
DCR Graphs: A decade of R&D
10
2007-11: Computer Supported Mobile Adaptive Business Processes
(Danish Research Foundation for Technology and Production)
2010: Case Studies of Best
Practice Workflow and
Workflow in Practice
(Innovation Network Project)
2008-2012: Trustworthy Pervasive Healthcare Processes
(Danish Strategic Research Council)
2009-13: Services in Context
(Danish Strategic Research Council)
32. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
DCR Graphs: A decade of R&D
10
2007-11: Computer Supported Mobile Adaptive Business Processes
(Danish Research Foundation for Technology and Production)
2010: Case Studies of Best
Practice Workflow and
Workflow in Practice
(Innovation Network Project)
2008-2012: Trustworthy Pervasive Healthcare Processes
(Danish Strategic Research Council)
2009-13: Services in Context
(Danish Strategic Research Council)
DCRGraphs.com
2011-2014: Flexible Cross-organizational Case Management
(Industrial PhD)
33. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
DCR Graphs: A decade of R&D
10
2007-11: Computer Supported Mobile Adaptive Business Processes
(Danish Research Foundation for Technology and Production)
2010: Case Studies of Best
Practice Workflow and
Workflow in Practice
(Innovation Network Project)
2008-2012: Trustworthy Pervasive Healthcare Processes
(Danish Strategic Research Council)
2009-13: Services in Context
(Danish Strategic Research Council)
2014-17: Computational Artifacts: Design Oriented Theory of
Computational Artifacts in Cooperative Work Practices
(Velux Foundation, www.COMPART.ku.dk)
DCR.Tools
DCRGraphs.com
2011-2014: Flexible Cross-organizational Case Management
(Industrial PhD)
34. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
DCR Graphs: A decade of R&D
10
2007-11: Computer Supported Mobile Adaptive Business Processes
(Danish Research Foundation for Technology and Production)
2010: Case Studies of Best
Practice Workflow and
Workflow in Practice
(Innovation Network Project)
2015: Finance Case:
Visualising for Customers, Business & IT
(Copenhagen Fintech Innovation & Research)
2008-2012: Trustworthy Pervasive Healthcare Processes
(Danish Strategic Research Council)
2009-13: Services in Context
(Danish Strategic Research Council)
2014-17: Computational Artifacts: Design Oriented Theory of
Computational Artifacts in Cooperative Work Practices
(Velux Foundation, www.COMPART.ku.dk)
DCR.Tools
DCRGraphs.com
2011-2014: Flexible Cross-organizational Case Management
(Industrial PhD)
35. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
DCR Graphs: A decade of R&D
10
2015-16: ProSec: Cyber security and ICT Infrastructure with importance to crucial functions
in Denmark - Mapping Emergency and Security Processes in the Danish Public Transport
Sector and their Dependency on ICT (Royal Danish Defence College)
2007-11: Computer Supported Mobile Adaptive Business Processes
(Danish Research Foundation for Technology and Production)
2010: Case Studies of Best
Practice Workflow and
Workflow in Practice
(Innovation Network Project)
2015: Finance Case:
Visualising for Customers, Business & IT
(Copenhagen Fintech Innovation & Research)
2008-2012: Trustworthy Pervasive Healthcare Processes
(Danish Strategic Research Council)
2009-13: Services in Context
(Danish Strategic Research Council)
2014-17: Computational Artifacts: Design Oriented Theory of
Computational Artifacts in Cooperative Work Practices
(Velux Foundation, www.COMPART.ku.dk)
DCR.Tools
DCRGraphs.com
2011-2014: Flexible Cross-organizational Case Management
(Industrial PhD)
36. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
DCRGraphs.com & DCR.Tools
11
Commercial tool
(free for test &
academia)
Academic tool
37. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Events & roles
12
Events as
“Post-it notes”
38. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Events & roles
12
Events as
“Post-it notes”
Can simulate at
any time
39. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Simulation
13
40. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Simulation
13
add computer as users
41. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Simulation
13
add computer as users
add friends as users
42. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Any task is possible
14
43. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Swim lane view of simulation
15
44. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Save simulation as Happy
16
45. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Write description
17
46. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Hurried case
18
47. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Hurried case not compliant
19
Fig. 10.1 Prespecified process model Smed
Table 10.1 Examples of compliance rules for medical processes
c1 Before a surgery may be performed the patient must be prepared for it and be sent to
the surgical suite.
c2 After examining the patient a decision must be made. However, this must not be done
before the examination.
c3 After the examination, the patient must be informed about the risks of the (planned)
surgery.
c4 Before scheduling the surgery the patient has to be informed about anesthesia.
c5 If a surgery has not been scheduled it must not be performed.
c6 After a patient is discharged a discharge letter must be written.
c7 After performing the surgery and before writing the discharge letter, a surgery report
must be created and a lab test be made.
particularly crucial for process instances defined or adapted on-the-fly (cf. Chap. 7),
i.e., for which there is no fully prespecified process model. Likewise, compliance
monitoring at run-time is required if a priori compliance checking is not feasible,
e.g., if the process model is too large or the compliance rules are too complex.
48. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Save unhappy hurried case
20
49. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Re-run or re-play simulations
21
re-run as test on current graph
re-play on
original graph
50. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Re-run or re-play simulations
22
51. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Add condition constraints
23
52. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Re-run simulations
24
53. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Conditions disable events
25
54. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Make decision enabled
26
55. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Also response constraints
27
Table 10.1 Examples of compliance rules for medical processes
c1 Before a surgery may be performed the patient must be prepared for it and be sent to
the surgical suite.
c2 After examining the patient a decision must be made. However, this must not be done
before the examination.
c3 After the examination, the patient must be informed about the risks of the (planned)
surgery.
c4 Before scheduling the surgery the patient has to be informed about anesthesia.
c5 If a surgery has not been scheduled it must not be performed.
c6 After a patient is discharged a discharge letter must be written.
c7 After performing the surgery and before writing the discharge letter, a surgery report
must be created and a lab test be made.
particularly crucial for process instances defined or adapted on-the-fly (cf. Chap. 7),
i.e., for which there is no fully prespecified process model. Likewise, compliance
monitoring at run-time is required if a priori compliance checking is not feasible,
e.g., if the process model is too large or the compliance rules are too complex.
Regarding completed process instances, in addition, a process-aware information
system (PAIS) needs to be able to determine whether these instances were executed
Responses
56. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Simulation with responses
28
required before the process can complete
57. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
No more pending responses
29
no more pending responses, process can end
58. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Excluded = irrelevant events
30
events may be excluded initially
http://www.dcrgraphs.net/Tool?id=4863#
59. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Dynamic inclusion & exclusion
31
events may be excluded initially
60. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk) 32
events may be excluded initially
Dynamic inclusion & exclusion
61. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Complete constraint map
33
http://www.dcrgraphs.net/Tool?id=4864#
62. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
DCR Graphs as formulars
34
63. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Data & Declarative Forms
35
Fig. 1. Reimbursement claim approval form (rejecting).
f the manager chooses “Approve”, the form is complete and may be sub-
mitted.
f the manager chooses “Reject”, a new field appears—the “Justify rejection”
ext box of Figure 1.
When this field appears, it must be filled in before the form can be submitted.
f the manager changes his mind and reverts his choice from “Reject” to
“Approve”, the description field should disappear again.
Fig. 2. Reimbursement claim approval form (initial appearance).
make two notes about this form and its behaviour.
irst, even this seemingly exceedingly simple form has fairly complex be-
our when you sit down and write it out as we did in the above list. Getting
behaviour right is not necessarily di cult, but it is time-consuming and
nsive because of the required programmer intervention.
econd, the form and the rules governing it are inextricably linked with the
ess to which the form contributes. The little list above is littered with men-
Fig. 1. Reimbursement claim approval form (rejecting).
3. If the manager chooses “Approve”, the form is complete and ma
mitted.
4. If the manager chooses “Reject”, a new field appears—the “Justify
text box of Figure 1.
5. When this field appears, it must be filled in before the form can be s
6. If the manager changes his mind and reverts his choice from “R
“Approve”, the description field should disappear again.
tions; this is an extension to the syntax and semantics of DCR graphs
has not yet appeared in the scientific literature.
DCR comprises (1) a set of activities and (2) a set of relations between
tivities. Activities are there to be executed, and relations indicate what
es happen to the state of the DCR graph as activities are executed. By
ntion, executing an activity in a DCR graph may input a data value for
tivity.
the name suggests a DCR graph is a graph: the nodes are activities, and
ges relations.
a running example, we will use the DCR graph depicted in Figure 3. This
graph is a minimal model of the travel reimbursement workflow sketched
tion 2. It has just two activities (boxes): Approve or reject (to the left) and
Rejection (to the right).
Fig. 3. Example DCR graph (reimbursement workflow)
*
*
*
64. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Data & Declarative Forms
35
Fig. 1. Reimbursement claim approval form (rejecting).
f the manager chooses “Approve”, the form is complete and may be sub-
mitted.
f the manager chooses “Reject”, a new field appears—the “Justify rejection”
ext box of Figure 1.
When this field appears, it must be filled in before the form can be submitted.
f the manager changes his mind and reverts his choice from “Reject” to
“Approve”, the description field should disappear again.
Fig. 2. Reimbursement claim approval form (initial appearance).
make two notes about this form and its behaviour.
irst, even this seemingly exceedingly simple form has fairly complex be-
our when you sit down and write it out as we did in the above list. Getting
behaviour right is not necessarily di cult, but it is time-consuming and
nsive because of the required programmer intervention.
econd, the form and the rules governing it are inextricably linked with the
ess to which the form contributes. The little list above is littered with men-
Fig. 1. Reimbursement claim approval form (rejecting).
3. If the manager chooses “Approve”, the form is complete and ma
mitted.
4. If the manager chooses “Reject”, a new field appears—the “Justify
text box of Figure 1.
5. When this field appears, it must be filled in before the form can be s
6. If the manager changes his mind and reverts his choice from “R
“Approve”, the description field should disappear again.
tions; this is an extension to the syntax and semantics of DCR graphs
has not yet appeared in the scientific literature.
DCR comprises (1) a set of activities and (2) a set of relations between
tivities. Activities are there to be executed, and relations indicate what
es happen to the state of the DCR graph as activities are executed. By
ntion, executing an activity in a DCR graph may input a data value for
tivity.
the name suggests a DCR graph is a graph: the nodes are activities, and
ges relations.
a running example, we will use the DCR graph depicted in Figure 3. This
graph is a minimal model of the travel reimbursement workflow sketched
tion 2. It has just two activities (boxes): Approve or reject (to the left) and
Rejection (to the right).
Fig. 3. Example DCR graph (reimbursement workflow)
*
*
*
ust have put data into it. In particular, the user is free to subsequently
validate the contents of the field.
eviate this tension, we must model more faithfully the dynamic be-
the form, taking into account the event of invalidating the required
do so by adding to our model a response arrow from the pending
itself, guarded by value = null. This means that whenever the field is
with a null value, it will be marked as pending again, ensuring that the
s not, in fact, complete before the activity/field is either excluded or
non-null value. Of course, one could define more complex criteria for
lue is invalid and thus trigger the response. In particular, the guard
end on the value entered in other fields.
4. Variant of DCR graph in Figure 3 taking null-values into account.
65. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Data & Declarative Forms
35
Forms are Declarative Processes! [BPM Case paper 2016]
Fig. 1. Reimbursement claim approval form (rejecting).
f the manager chooses “Approve”, the form is complete and may be sub-
mitted.
f the manager chooses “Reject”, a new field appears—the “Justify rejection”
ext box of Figure 1.
When this field appears, it must be filled in before the form can be submitted.
f the manager changes his mind and reverts his choice from “Reject” to
“Approve”, the description field should disappear again.
Fig. 2. Reimbursement claim approval form (initial appearance).
make two notes about this form and its behaviour.
irst, even this seemingly exceedingly simple form has fairly complex be-
our when you sit down and write it out as we did in the above list. Getting
behaviour right is not necessarily di cult, but it is time-consuming and
nsive because of the required programmer intervention.
econd, the form and the rules governing it are inextricably linked with the
ess to which the form contributes. The little list above is littered with men-
Fig. 1. Reimbursement claim approval form (rejecting).
3. If the manager chooses “Approve”, the form is complete and ma
mitted.
4. If the manager chooses “Reject”, a new field appears—the “Justify
text box of Figure 1.
5. When this field appears, it must be filled in before the form can be s
6. If the manager changes his mind and reverts his choice from “R
“Approve”, the description field should disappear again.
tions; this is an extension to the syntax and semantics of DCR graphs
has not yet appeared in the scientific literature.
DCR comprises (1) a set of activities and (2) a set of relations between
tivities. Activities are there to be executed, and relations indicate what
es happen to the state of the DCR graph as activities are executed. By
ntion, executing an activity in a DCR graph may input a data value for
tivity.
the name suggests a DCR graph is a graph: the nodes are activities, and
ges relations.
a running example, we will use the DCR graph depicted in Figure 3. This
graph is a minimal model of the travel reimbursement workflow sketched
tion 2. It has just two activities (boxes): Approve or reject (to the left) and
Rejection (to the right).
Fig. 3. Example DCR graph (reimbursement workflow)
*
*
*
ust have put data into it. In particular, the user is free to subsequently
validate the contents of the field.
eviate this tension, we must model more faithfully the dynamic be-
the form, taking into account the event of invalidating the required
do so by adding to our model a response arrow from the pending
itself, guarded by value = null. This means that whenever the field is
with a null value, it will be marked as pending again, ensuring that the
s not, in fact, complete before the activity/field is either excluded or
non-null value. Of course, one could define more complex criteria for
lue is invalid and thus trigger the response. In particular, the guard
end on the value entered in other fields.
4. Variant of DCR graph in Figure 3 taking null-values into account.
67. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Modelling & Validation
37
Papers: [ACM15,BPM13-15]
Papers: [BPM14]
68. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
DCR Graphs for execution
38
Workflow engine
Papers: [ACM15,BPM13-15]
69. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Seamless & safe distribution
39
Workflow engine Workflow engine
Papers: [SEFM2011,FHIES2011,BPM15]
70. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Monitoring & Compliance
40
Workflow engine Workflow engine
Run-time monitor
Papers: [R. Mukkamala PhD,CSF 2016]
71. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Policy enforcement
41
Workflow engine Workflow engine
Policy enforcement point
Papers: [CSF 2016]
72. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
What is special for DCR graphs?
42
73. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
What is special for DCR graphs?
• Formal and close to natural language:
Conditions, Responses, Inclusions and Exclusions
42
74. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
What is special for DCR graphs?
• Formal and close to natural language:
Conditions, Responses, Inclusions and Exclusions
• Expressive and decidable:
Can express all regular safety and liveness properties
42
75. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
What is special for DCR graphs?
• Formal and close to natural language:
Conditions, Responses, Inclusions and Exclusions
• Expressive and decidable:
Can express all regular safety and liveness properties
• Operational and understandable:
Run-time state as “check-list” on events
42
76. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
What is special for DCR graphs?
• Formal and close to natural language:
Conditions, Responses, Inclusions and Exclusions
• Expressive and decidable:
Can express all regular safety and liveness properties
• Operational and understandable:
Run-time state as “check-list” on events
• Efficient distributed monitoring & enactment
Locally determine enabledness & effect of events
42
77. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Additional work
• Time & Dynamic Subprocesses
• Independence analysis (true concurrency)
• Applications to case studies [FHIES2011,ACM14,BPM15,S4CIP16]
(Healthcare, funding agency, finance & emergency management)
• Run-time adaptation & refinement
[EDOC2013][ACM14][FM15]
• Programming Language, Data & Forms
43
[JLAP82,2013, BPM14,FM15]
[DEBS2012,REBLS15,
BPM CASE 2016]
[BPM15]
78. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Dealing with state space explosion
• Safe distribution & infer concurrency
• Modularity & step-wise refinement
• Static analysis for reachability & enforceability [CSF15]
44
[SEFM2011,BPM15]
[FM15]
vs
79. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
True concurrency
45
Provides an asynchronous transition system
80. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Timed DCR Graphs
46
Eventually is often not good enough….
81. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Timed DCR Graphs
46
Eventually is often not good enough….
and delays may be required
82. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Timed DCR Graphs
46
Eventually is often not good enough….
and delays may be required
Timed DCR Graphs introduce
delays on conditions, and deadlines on responses
[JLAP82,2013,
CSF2016]
83. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Delays & Deadlines
47
“After eating you must wait 12 hours before surgery”
“After surgery, a checkup must be done within 7 days”
84. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Enforceability & Escalation
48
Some events are uncontrollable
in particular progress of time and human activities
85. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Enforceability & Escalation
48
Some events are uncontrollable
in particular progress of time and human activities
Need compensation/escalation &
pro-active enforcement [CSF2016]
86. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Enforceability & Escalation
48
Some events are uncontrollable
in particular progress of time and human activities
Need compensation/escalation &
pro-active enforcement [CSF2016]
Try research-prototype at dcr.tools/obligations
87. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Sub processes & infinity
• Dynamically spawned sub processes
• supported in theory and tools - but makes
termination undecidable
• not a problem for enactment, run-time monitoring &
(some) static analysis (refinement)
49
[FM15]
[FM15]
88. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Refinement & Security
50
89. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Refinement & Security
50
(EU General
Data Protection
Requirement)
90. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Mining Declarative Processes
51
Process log Statistically inferred
constraints
Inferred flows
Avoid mined spaghetti diagrams
91. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
First DCR mining algorithm
52
Contradiction based mining:
1) Start with every possible
condition, response, exclusion
and every activity excluded,
pending & not executed
2) Remove relation & include
activity when contradiction
found in log
Declarative Process Mining for DCR Graphs⇤
Søren Debois
IT University of Copenhagen
Copenhagen, Denmark
debois@itu.dk
Thomas T. Hildebrandt
IT University of Copenhagen
Copenhagen, Denmark
hilde@itu.dk
Paw Høvsgaard Laursen
IT University of Copenhagen
Copenhagen, Denmark
pawh@itu.dk
Kenneth Ry Ulrik
IT University of Copenhagen
Copenhagen, Denmark
kulr@itu.dk
ABSTRACT
We investigate process mining for the declarative Dynamic
Condition Response (DCR) graphs process modelling lan-
guage. We contribute (a) a process mining algorithm for
DCR graphs, (b) a proposal for a set of metrics quantifying
output model quality, and (c) a preliminary example-based
comparison with the Declare Maps Miner. The algorithm
takes a contradiction-based approach, that is, we initially
assume that all possible constraints hold, subsequently re-
moving constraints as they are observed to be violated by
traces in the input log.
Keywords
Declarative process mining; DCR graphs
1. INTRODUCTION
Business process management (BPM) technologies [33]
support the management and digitalisation of workflows and
business processes by employing explicit process models, fol-
lowing a cycle of process (re)design, validation, execution
and monitoring.
Process mining algorithms [32] have been proposed for the
identification of process models from process logs, support-
ing both process design and compliance monitoring.
Most industrial BPM tools and process miners describe
processes as imperative flow diagrams such as BPMN. How-
ever, flow diagrams tend to get either too rigid or too com-
plex, in particular for knowledge work processes having a
high degree of variation [28]. Moreover, flow diagrams only
describe how to perform a process, leaving a gap to the legal
regulations and guidelines, that are often more declarative
in nature, describing why the process must be performed in
⇤Authors listed alphabetically. This work supported in part
by the Velux Foundation, grant 33295, and Exformatics
A/S.
Permission to make digital or hard copies of part or all of this work for personal or
classroom use is granted without fee provided that copies are not made or distributed
for profit or commercial advantage and that copies bear this notice and the full citation
on the first page. Copyrights for third-party components of this work must be honored.
For all other uses, contact the owner/author(s).
SAC 2017 April 03-07, 2017, Marrakech, Morocco
c 2017 Copyright held by the owner/author(s).
ACM ISBN 978-1-4503-4486-9/17/04.
DOI: http://dx.doi.org/10.1145/3019612.3019622
certain ways, not how exactly it must be performed. For
instance, a clinical guideline may state, that a patient must
consent to a blood transfusion [13]. It does not state ex-
actly when such consent should be obtained, only “prior to
the transfusion”.
For this reason, it is recommended to use flow diagrams
only for routine processes, or for describing common stan-
dard practices and allow deviations [28]. It has been advo-
cated that declarative notations should be used as output of
process mining (e.g. [17]) and for run-time process support
(e.g. [25, 24, 29]). For the former, one hopes to extract
from a process log the rules obeyed in practice (the “why”)
as opposed to a flow-diagram describing the usual executions
(the “how”). For the latter, one hopes to guide knowledge
workers to activities in conformance with rules and regula-
tions.
Implementation techniques for most declarative models
such as Declare [27] and DecSerFlow [31], rely on translating
the declarative constraints to an imperative model (e.g., an
automaton [20]) to enable execution. Such translation usu-
ally entail a state-space explosion, and run-time adaptation
of constraints becomes more di cult, because the automa-
ton must be recomputed when constraints change.
A notable exception is the Dynamic Condition Response
(DCR) graphs process language [11, 30]. DCR graphs can be
executed without intermediate transformation to an imper-
ative model creating the entire transition graph, and more
directly support run-time adaptive case management [24,
5]. DCR graphs are supported by industrial design and case
management tools (see e.g. dcrgraphs.net and [5]).
In the present paper, we present the first process mining
algorithm for DCR graphs.
2. DCR GRAPHS
In this Section, we briefly recall DCR graphs. For a formal
introduction and applications, refer to [11, 23, 30, 3, 5, 6].
Dynamic Condition Response graphs is a declarative mod-
elling notation describing at the same time a process and
its run-time state. The core notation comprises activities,
activity states, and four relations between activities. An ac-
tivity state comprises three booleans, indicating respectively
whether the activity has been executed, is included, and
is pending. Intuitively, activities that are not included are
treated as temporarily absent from the workflow; activities
that are pending must eventually be executed or excluded
before the workflow may complete.
[SAC 2017]
Implemented at dcr.tools
92. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Example
53
Provide Budget
Calculate Maximum Cost
Find Houses
Buy
Provide Budget
Calculate Maximum Cost
Find Houses
Logs
93. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Remove redundancy
54
Provide Budget
Calculate Maximum Cost
Find Houses
Buy
Provide Budget
Calculate Maximum Cost
Find Houses
Logs
94. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Add trace & remove redundancy
55
Provide Budget
Calculate Maximum Cost
Find Houses
Buy
Provide Budget
Calculate Maximum Cost
Find Houses
Provide Budget
Calculate Maximum Cost
Find Houses
Provide Budget
Logs
ProvideBudget CalculateMaximumCost FindHouses
http://dcr.itu.dk/Workbench/Default/2265168991/1033131530
95. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Not always perfect fit
56
http://dcr.itu.dk/Workbench/Default/2489849454/4066076145
96. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Can view as transition system
57
97. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
What is the new LTS?
58
http://dcr.itu.dk/Workbench/Default/578939190/4034201725
98. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
What is the new LTS?
59
http://dcr.itu.dk/Workbench/Default/578939190/4034201725
99. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Thoughts on DCR process Miner
• DCR process miner heavily uses include/exclude
• Future work:
• Most frequent activities included initially ?
• Parallelise (seems promising)
• Time? Map to true concurrent model
We collaborate with KMD and danish municipalities.
Look out for possibilities to join our team as postdoc -
write me if interested.
60
100. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
True Concurrent Mining
61
101. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Conclusions
62
102. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Conclusions
• Flow-graphs take routes as starting point: Often too
inflexible, do not capture why, difficult to maintain
62
103. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Conclusions
• Flow-graphs take routes as starting point: Often too
inflexible, do not capture why, difficult to maintain
• DCR graphs take events and constraints as starting
point: Support flexibility, adaptability and formal
validation of compliance and correctness
Can express all omega-regular concurrent languages
62
104. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Conclusions
• Flow-graphs take routes as starting point: Often too
inflexible, do not capture why, difficult to maintain
• DCR graphs take events and constraints as starting
point: Support flexibility, adaptability and formal
validation of compliance and correctness
Can express all omega-regular concurrent languages
• Tool support & applied with success in industry
62
105. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Conclusions
• Flow-graphs take routes as starting point: Often too
inflexible, do not capture why, difficult to maintain
• DCR graphs take events and constraints as starting
point: Support flexibility, adaptability and formal
validation of compliance and correctness
Can express all omega-regular concurrent languages
• Tool support & applied with success in industry
• Early work on mining [SAC17] and refinement [FM15]
62
106. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
Conclusions
• Flow-graphs take routes as starting point: Often too
inflexible, do not capture why, difficult to maintain
• DCR graphs take events and constraints as starting
point: Support flexibility, adaptability and formal
validation of compliance and correctness
Can express all omega-regular concurrent languages
• Tool support & applied with success in industry
• Early work on mining [SAC17] and refinement [FM15]
• Look out for postdoc & ph.d. positions to join us!
62
107. IT UNIVERSITY OF COPENHAGEN
Modelling & Mining Event-based Concurrent Declarative Processes as DCR Graphs ATAED 26/6/17
Thomas T. Hildebrandt (hilde@itu.dk)
The future
63
S (Italy)
d
gi,
m
Municipality partners
as early adopters:
Koncern IT - Copenhagen Municipality
IT & Digitalisation,
Syddjurs Municipality
Kammeradvokaten
& Globeteam
ve case management for knowledge workers
ation of knowledge work processes
regulations (e.g. data protection)
ness and legal compliance
Enabling technologies
shared as open source tools
via the OS2 open source
digitalisation community
Effective, co-created & compliant
adaptive case management solutions
for knowledge workers
decision
support
enter
• Lov om Sygedagpenge
(LOV nr 563 af 09/06/2006)
• Retssikkerhedsloven
(LBK nr 1054 af 07/09/2010)
• Datagrundlag
(BEK nr 418 af 23/04/2010)
of log-files
Process-mining
Pro-active compliance
Prescriptive BPM