Business Process Simulation 2II75

Download Report

Transcript Business Process Simulation 2II75

Models vs. Reality
dr.ir. B.F. van Dongen
Assistant Professor
Eindhoven University of Technology
[email protected]
Process Mining
• Discovering processes
• How do people behave?
• Compliance oriented
• Where and why do people
deviate from standards / rules /
regulations?
• Performance oriented
• Where are bottlenecks
in my processes?
Aligning models to Observed Behavior
• Starting point for conformance checking is a process model
and a log
Introduction: Alignments
• Alignments are used for conformance checking
• Alignments are computed over a trace and a model:
− A trace is a (partial) order of activities
− A model is a labeled Petri labeled with activities
• An alignment explains exactly where deviations occur:
− A synchronous move mean that an activity is in the log and a corresponding
transition was enabled in the model
− A log move means that no corresponding activity is found in the model
− A model move means that no corresponding activity appeared in the log
Example
model:
ABDE
…
…
log
Logged “A” aligns nicely to model
model:
ABDE
…
…
log
A
A
Logged “B” aligns nicely to model
model:
ABDE
…
…
log
A B
A B
Logged “D” does not fit the model
model:
ABDE
…
…
log
A B
A B D
“C” was probably executed, but was not logged
model:
ABDE
…
…
log
C
A B
A B D
Logged “E” aligns nicely to model
model:
ABDE
…
…
log
C E
A B
E
A B D
Alignment shows where deviations occurred
Alignment:
The best way to fit
the trace in the
model
model:
ABDE
…
…
log
C E
A B
E
A B D
Alignments
• Alignments specify exactly where deviations
occurred when comparing logs to models
• Alignments can be used for:
•
•
•
•
•
Fitness/precision computations
Performance analysis
Model repair
...
Compliance analysis
Use of alignment techniques in compliance
?
compliance improvement
compliance checking and analysis
implement compliance measures
formalize compliance rules
elicit compliance rules
13
Automated compliance checking
business process
compliance
requirement
compliance
diagnostic
information
specification
compliance checker
Automated compliance checking
business process
Log
compliance
requirement
B
A
F
B
Ƭ
F
compliance
Petri net pattern
alignment
compliance checker
diagnostic
information
Specifying Compliance Rules
Which
compliance
pattern?
How to
prune the
Petri net
pattern?
rule repository
Log
compliance
checker
precise
Petri net pattern
compliance specifier
Implementation
ProM6 (www.promtools.org/prom6) Compliance
Elicit Compliance
Checking
RuleUsing Conformance Checking
Patient registration
X-ray
Patient
registration
Patient
X-Ray registration
others
Conclusions
• Alignments provide a powerful method to explain where
operational processes deviated from models
• Using the right models, alignments can detect (and predict)
possible violations of compliance rules
• Alignments provide guarantees on non-deviating cases
Future directions
Current challenges:
1. Representation and extraction of multi-dimensional event data
for deviation detection
2. Representation and management of deviations
3. Detection and diagnosis of deviations
4. Online, real time deviation prediction
5. Integration of prototypes applicable to high-volume data
6. Application on real-life cases
Questions