Temporal Compliance Checking
CAiSE‟2013
Elham Ramezani
Dirk Fahland
Boudewijn van Dongen
Wil M.P. van der Aalst
Compliance
Report
✓
✗
✓
✗
1
Enterprises are governed by regulations
Laws & Regulations
Diagnostic
Information
2
Compliance management and BPM
Compliance checking
on log
Compliance checking and process
perspectives
3
Data Perspective
Resource Perspective
Time Perspective
Activity „A‟ must occur at time t.
Capture occurrence of activity „A‟.
Control-Flow
Compliance Checking
Temporal Compliance
Checking
Compliance
CheckerRule 3
Rule 2
Problem setting
4
Rule 1
Diagnostic
Information
CF-Rule
Event Log
Compliance
Requirement
T-Rule
Example
5
A certain antibiotic should be administered to the patient on cycles of 2-time
treatment. The administration of antibiotic can be extended for another cycle , but
there should be at least a week delay between two subsequent cycles.
A : antibiotic administration
… : other activities in the trace
7d
Sample trace:
• C-F Rule: Activity ‘A’ must occur in cycles of 2-time occurrences.
A … A … A … A … AA,1 …,1 A,2 …,3 A,9 …,10 A,14 …,15 A,15
• T-Rule: There must be 7 days delay between two subsequent cycles.
Multiple possible diagnostics
6
A,1 …,1 A,2 …,3 A,9 …,10 A,14 …,15 A,15
A … A … A A … A … A
Log
3rd. possibility:
Log
2nd. possibility:
Log
1st. possibility:
4th.
Sample trace:
A,1 …,1 A,2 …,3 A,9 …,10 A,14 …,15 A,15
A … A … A … A A … A
A,1 …,1 A,2 …,3 A,9 …,10 A,14 …,15 A,15
A A … A … A … A … A
A,1 …,1 A,2 …,3 A,9 …,10 A,14 …,15 A,15
A is missing
Enriched log +
Diagnostics
Rule 3
Rule 2
… solve by Conformance Checking
7
Rule 1
Diagnostic
Information
Event
Log
Compliance
Requirement
Petri net
Specification
Conformance
Checker
Data-Aware
Petri net
Specification
CF-Rule
T-Rule
Data-Aware
Conformance
Checker
Compliance
Checker
Enriched log +
Diagnostics
Rule 3
Rule 2
8
Rule 1
Diagnostic
Information
Event
Log
Compliance
Requirement
Petri net
Specification
Conformance
Checker
Data-Aware
Petri net
Specification
CF-Rule
T-Rule
Data-Aware
Conformance
Checker
Compliance
Checker
Control-flow alignment
9
Alignment  Diagnostic information
trace:
alignment: actual trace vs. correct trace
“move on model” = event is missing invisible
correct
Event
Log
Adriansyah et al.[3]
Conformance
Checker
A,1 A,2 A,9 A,14 A,15
2nd.alignment
3rd.
S Ist A A Icmp Ist A A Icmp Ist A A Icmp E
>> >> A,
1
A,
2
>> >> A,
9
>> >> >> A,
14
A,
15
>> >>
Enriched log +
Diagnostics
Rule 3
Rule 2
10
Rule 1
Diagnostic
Information
Event
Log
Compliance
Requirement
Petri net
Specification
Conformance
Checker
Data-Aware
Petri net
Specification
CF-Rule
T-Rule
Data-Aware
Conformance
Checker
Compliance
Checker
Log Enrichment
11
Enriching log with additional information
S Ist A,
1
A,
2
Icmp Ist A,
9
A Icmp Ist A,
14
A,
15
Icmp E
>> >> A,
1
A,
2
>> >> A,
9
>> >> >> A,
14
A,
15
>> >>
S,
1
Ist,
1
A,
1
A,
2
Icmp,
2
Ist,
9
A,
9
A
9
Icmp,
9
Ist,
14
A,
14
A,
15
Icmp,
15
E,
15
Enriched log from
2nd.alignment
3rd. 9
Enriched log +
Diagnostics
Rule 3
Rule 2
12
Rule 1
Diagnostic
Information
Event
Log
Compliance
Requirement
Petri net
Specification
Conformance
Checker
Data-Aware
Petri net
Specification
CF-Rule
T-Rule
Data-Aware
Conformance
Checker
Compliance
Checker
Temporal rule formalization
13
Between two cycles of occurrences of ‘A’ there must be one week delay
…
EFinal
Initial
…
Ist
Icmp
A
S
[tst` = tcmp+ 7d] ∨ tcmp = undef.
Formalizing temporal rule in terms of
data-aware Petri net
S,
1
Ist,
1
A,
1
A,
2
Icmp,
2
Ist,
9
A,
9
A
9
Icmp,
9
Ist,
14
A,
14
A,
15
Icmp,
15
E,
15
Enriched log:
W: { t st }
W: { t cmp }
t st =1 t A =1 t A =2 t cmp =2t st =9 t A =9
✓
t A =9t cmp =9 t st =14
✗
t A =14t A =15t cmp =15t S=1 t E =15
Enriched log +
Diagnostics
Rule 3
Rule 2
14
Rule 1
Diagnostic
Information
Event
Log
Compliance
Requirement
Petri net
Specification
Conformance
Checker
Data-Aware
Petri net
Specification
CF-Rule
T-Rule
Data-Aware
Conformance
Checker
Compliance
Checker
Temporal compliance checking
PAGE 15
Data-aware alignment
Data-Aware
Conformance Checker
M.de Leoni et al. [13]
S,
1
Ist,
1
A,
1
A,
2
Icmp,
2
Ist,
9
A,
9
A,
9
Icmp,
9
Ist,
16
A,
14
A,
15
Icmp,
15
E,
15
S,
1
Ist,
1
A,
1
A,
2
Icmp,
2
Ist,
9
A,
9
A,
9
Icmp,
9
Ist,
14
A,
14
A,
15
Icmp,
15
E,
15
✗
L: 14, M: 16
✓✓
PAGE 16
Diagnostics
Data-Aware
Conformance Checker
Conformance Checker
S,
1
Ist,
1
A,
1
A,
2
Icmp,
2
Ist,
9
A,
9
A,
9
Icmp,
9
Ist,
16
A,
14
A,
15
Icmp,
15
E,
15
S,
1
Ist,
1
A,
1
A,
2
Icmp,
2
Ist,
9
A,
9
A,
9
Icmp,
9
Ist,
14
A,
14
A,
15
Icmp,
15
E,
15
Combined
Diagnostic Information
PAGE 17
Implementation
ProM6 (www.promtools.org/prom6)
Check Temporal Compliance Using Data-Aware Conformance Checking
18
Case study in Dutch municipalities
Compliance Constraint:
Applications must be processed within 8 weeks, excluding the time they
are waiting for additional required information provided by applicant.
# Cases
Violations
# Avg. Delay
(months)
Max. Delay
(months)
1364 233 3.5 15
19
Cases vs. handover of work
0
10
20
30
40
50
60
70
0 10 20 30 40 50 60 70
# cases
# handovers
20
Violations vs. handover of work
# handovers
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
25 30 35 40 45 50 55 60 65 70
% violated
% violated
Enriched log +
Diagnostics
Rule 3
Rule 2
21
Rule 1
Diagnostic
Information
Event
Log
Compliance
Requirement
Petri net
Specification
Conformance
Checker
Data-Aware
Petri net
Specification
CF-Rule
T-Rule
Data-Aware
Conformance
Checker
Compliance
Checker
Take home points
22
Questions?
Enriched log +
Diagnostics
Rule 3
Rule 2
23
Rule 1
Diagnostic
Information
Event
Log
Compliance
Requirement
Petri net
Specification
Conformance
Checker
Data-Aware
Petri net
Specification
CF-Rule
T-Rule
Data-Aware
Conformance
Checker
Compliance
Checker
Rule repositories
Repository
Generic CF-Rules
& Formalization
Repository
Generic T-Rules &
Formalization
CF-Rule
T-Rule
Overview
24
Thank you for your
attention!

Ramezani taghiabadi temporal compliance checking 2

  • 1.
    Temporal Compliance Checking CAiSE‟2013 ElhamRamezani Dirk Fahland Boudewijn van Dongen Wil M.P. van der Aalst
  • 2.
    Compliance Report ✓ ✗ ✓ ✗ 1 Enterprises are governedby regulations Laws & Regulations Diagnostic Information
  • 3.
    2 Compliance management andBPM Compliance checking on log
  • 4.
    Compliance checking andprocess perspectives 3 Data Perspective Resource Perspective Time Perspective Activity „A‟ must occur at time t. Capture occurrence of activity „A‟. Control-Flow Compliance Checking Temporal Compliance Checking
  • 5.
    Compliance CheckerRule 3 Rule 2 Problemsetting 4 Rule 1 Diagnostic Information CF-Rule Event Log Compliance Requirement T-Rule
  • 6.
    Example 5 A certain antibioticshould be administered to the patient on cycles of 2-time treatment. The administration of antibiotic can be extended for another cycle , but there should be at least a week delay between two subsequent cycles. A : antibiotic administration … : other activities in the trace 7d Sample trace: • C-F Rule: Activity ‘A’ must occur in cycles of 2-time occurrences. A … A … A … A … AA,1 …,1 A,2 …,3 A,9 …,10 A,14 …,15 A,15 • T-Rule: There must be 7 days delay between two subsequent cycles.
  • 7.
    Multiple possible diagnostics 6 A,1…,1 A,2 …,3 A,9 …,10 A,14 …,15 A,15 A … A … A A … A … A Log 3rd. possibility: Log 2nd. possibility: Log 1st. possibility: 4th. Sample trace: A,1 …,1 A,2 …,3 A,9 …,10 A,14 …,15 A,15 A … A … A … A A … A A,1 …,1 A,2 …,3 A,9 …,10 A,14 …,15 A,15 A A … A … A … A … A A,1 …,1 A,2 …,3 A,9 …,10 A,14 …,15 A,15 A is missing
  • 8.
    Enriched log + Diagnostics Rule3 Rule 2 … solve by Conformance Checking 7 Rule 1 Diagnostic Information Event Log Compliance Requirement Petri net Specification Conformance Checker Data-Aware Petri net Specification CF-Rule T-Rule Data-Aware Conformance Checker Compliance Checker
  • 9.
    Enriched log + Diagnostics Rule3 Rule 2 8 Rule 1 Diagnostic Information Event Log Compliance Requirement Petri net Specification Conformance Checker Data-Aware Petri net Specification CF-Rule T-Rule Data-Aware Conformance Checker Compliance Checker Control-flow alignment
  • 10.
    9 Alignment  Diagnosticinformation trace: alignment: actual trace vs. correct trace “move on model” = event is missing invisible correct Event Log Adriansyah et al.[3] Conformance Checker A,1 A,2 A,9 A,14 A,15 2nd.alignment 3rd. S Ist A A Icmp Ist A A Icmp Ist A A Icmp E >> >> A, 1 A, 2 >> >> A, 9 >> >> >> A, 14 A, 15 >> >>
  • 11.
    Enriched log + Diagnostics Rule3 Rule 2 10 Rule 1 Diagnostic Information Event Log Compliance Requirement Petri net Specification Conformance Checker Data-Aware Petri net Specification CF-Rule T-Rule Data-Aware Conformance Checker Compliance Checker Log Enrichment
  • 12.
    11 Enriching log withadditional information S Ist A, 1 A, 2 Icmp Ist A, 9 A Icmp Ist A, 14 A, 15 Icmp E >> >> A, 1 A, 2 >> >> A, 9 >> >> >> A, 14 A, 15 >> >> S, 1 Ist, 1 A, 1 A, 2 Icmp, 2 Ist, 9 A, 9 A 9 Icmp, 9 Ist, 14 A, 14 A, 15 Icmp, 15 E, 15 Enriched log from 2nd.alignment 3rd. 9
  • 13.
    Enriched log + Diagnostics Rule3 Rule 2 12 Rule 1 Diagnostic Information Event Log Compliance Requirement Petri net Specification Conformance Checker Data-Aware Petri net Specification CF-Rule T-Rule Data-Aware Conformance Checker Compliance Checker Temporal rule formalization
  • 14.
    13 Between two cyclesof occurrences of ‘A’ there must be one week delay … EFinal Initial … Ist Icmp A S [tst` = tcmp+ 7d] ∨ tcmp = undef. Formalizing temporal rule in terms of data-aware Petri net S, 1 Ist, 1 A, 1 A, 2 Icmp, 2 Ist, 9 A, 9 A 9 Icmp, 9 Ist, 14 A, 14 A, 15 Icmp, 15 E, 15 Enriched log: W: { t st } W: { t cmp } t st =1 t A =1 t A =2 t cmp =2t st =9 t A =9 ✓ t A =9t cmp =9 t st =14 ✗ t A =14t A =15t cmp =15t S=1 t E =15
  • 15.
    Enriched log + Diagnostics Rule3 Rule 2 14 Rule 1 Diagnostic Information Event Log Compliance Requirement Petri net Specification Conformance Checker Data-Aware Petri net Specification CF-Rule T-Rule Data-Aware Conformance Checker Compliance Checker Temporal compliance checking
  • 16.
    PAGE 15 Data-aware alignment Data-Aware ConformanceChecker M.de Leoni et al. [13] S, 1 Ist, 1 A, 1 A, 2 Icmp, 2 Ist, 9 A, 9 A, 9 Icmp, 9 Ist, 16 A, 14 A, 15 Icmp, 15 E, 15 S, 1 Ist, 1 A, 1 A, 2 Icmp, 2 Ist, 9 A, 9 A, 9 Icmp, 9 Ist, 14 A, 14 A, 15 Icmp, 15 E, 15 ✗ L: 14, M: 16 ✓✓
  • 17.
    PAGE 16 Diagnostics Data-Aware Conformance Checker ConformanceChecker S, 1 Ist, 1 A, 1 A, 2 Icmp, 2 Ist, 9 A, 9 A, 9 Icmp, 9 Ist, 16 A, 14 A, 15 Icmp, 15 E, 15 S, 1 Ist, 1 A, 1 A, 2 Icmp, 2 Ist, 9 A, 9 A, 9 Icmp, 9 Ist, 14 A, 14 A, 15 Icmp, 15 E, 15 Combined Diagnostic Information
  • 18.
    PAGE 17 Implementation ProM6 (www.promtools.org/prom6) CheckTemporal Compliance Using Data-Aware Conformance Checking
  • 19.
    18 Case study inDutch municipalities Compliance Constraint: Applications must be processed within 8 weeks, excluding the time they are waiting for additional required information provided by applicant. # Cases Violations # Avg. Delay (months) Max. Delay (months) 1364 233 3.5 15
  • 20.
    19 Cases vs. handoverof work 0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 70 # cases # handovers
  • 21.
    20 Violations vs. handoverof work # handovers 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 25 30 35 40 45 50 55 60 65 70 % violated % violated
  • 22.
    Enriched log + Diagnostics Rule3 Rule 2 21 Rule 1 Diagnostic Information Event Log Compliance Requirement Petri net Specification Conformance Checker Data-Aware Petri net Specification CF-Rule T-Rule Data-Aware Conformance Checker Compliance Checker Take home points
  • 23.
  • 24.
    Enriched log + Diagnostics Rule3 Rule 2 23 Rule 1 Diagnostic Information Event Log Compliance Requirement Petri net Specification Conformance Checker Data-Aware Petri net Specification CF-Rule T-Rule Data-Aware Conformance Checker Compliance Checker Rule repositories Repository Generic CF-Rules & Formalization Repository Generic T-Rules & Formalization CF-Rule T-Rule Overview
  • 25.
    24 Thank you foryour attention!

Editor's Notes

  • #7 Where exactly sth is missing and when did it go wrong?