1. Dealing with Changes of Time-Aware Processes
Andreas Lanz, Manfred Reichert
Dealing with Changes of Time-Aware Processes
Image by saavem

2. Motivation Time and temporal restrictions are omnipresent Durations
Dealing with Changes of Time-Aware Processes | Andreas Lanz, Manfred Reichert |
Motivation
Time and temporal restrictions are omnipresent
Durations
Appointments
Deadlines …
Time perspective raises fundamental challenges for any process-aware information system
Existing PAIS offer only a limited support for modeling and managing time-aware processes [Lanz et al., 14]

3. Motivation The Challenge: Time can neither be slowed down nor stopped
Dealing with Changes of Time-Aware Processes | Andreas Lanz, Manfred Reichert |
Motivation
The Challenge:
Time can neither be slowed down nor stopped
Time-aware processes need to be flexible to cope with unforeseen events or delays
Should be possible to dynamically adapt time-aware process instances
Re-schedule deadlines, dynamically modify temporal constraints
Structurally change a process instance
“Lost time is never found again.”
– Benjamin Franklin ( )

4. Temporal Constraint Changes
Dealing with Changes of Time-Aware Processes | Andreas Lanz, Manfred Reichert |
Motivation
Change operations for time-aware processes
Temporal Consistency of the process instance needs to be ensured
Control Flow Changes
Insert Activity Serially
Insert Activity Parallel
Insert Activity Conditionally
Delete Activity
Temporal Constraint Changes
Insert Time Lag
Insert Fixed Date Element
Delete Time Lag
Delete Fixed Date Element

5. Time-Aware Processes Motivation
Dealing with Changes of Time-Aware Processes | Andreas Lanz, Manfred Reichert |
Time-Aware Processes
Motivation
Change Operations for Time-Aware Processes
Analyzing the Effects
Conclusion

6. Time-aware Processes – Time Constraints*
Dealing with Changes of Time-Aware Processes | Andreas Lanz, Manfred Reichert |
Time-aware Processes – Time Constraints*
Time Constraints
Activity Durations
Time Lags (min waiting time, max delay, …)
Fixed Date Elements (appointments, deadlines…)
A 𝑚𝑖𝑛𝐷, 𝑚𝑎𝑥𝐷
Minimum
Maximum A B
𝑆 𝑚𝑖𝑛𝐷,𝑚𝑎𝑥𝐷 𝐸
Start-Start
Start-End
End-Start
End-End
Latest Finish Date A
Earliest Start
Latest Start
Earliest Completion
Latest Completion
* [Lanz et al., 14]

7. Modeling Time-Aware Processes
Dealing with Changes of Time-Aware Processes | Andreas Lanz, Manfred Reichert |
Modeling Time-Aware Processes
Important Question:
Can an instance of the process model be completed without violating any time constraints?
 temporal consistency
Solution:
Conditional Simple Temporal Network (CSTN)*
= Simple Temporal Network + Propositions
* [Tsamardinos et al., 03, Hunsberger et al., 12]

8. Conditional Simple Temporal Network* (CSTN)
Dealing with Changes of Time-Aware Processes | Andreas Lanz, Manfred Reichert |
Conditional Simple Temporal Network* (CSTN)
Time Constraint
constraint
time point A B
𝑥,𝑦 ,𝛽
𝑥≤𝐵−𝐴≤𝑦 =𝐵−𝐴∈[𝑥,𝑦]
proposition representing the execution path
Consistency Check*
May restrict existing constraints
Derives interdependencies A B
𝑥 ′ , 𝑦 ′ , 𝛽
𝑦 ′ ≤𝑦
𝑥′≥𝑥
𝑥′≤𝐵−𝐴≤𝑦′
* [Tsamardinos et al., 03, Hunsberger et al., 12]

9. CSTN Transformation* A 𝑥, 𝑦 AS AE XS XE X Activity 0,∞ , 𝛽 𝑥,𝑦 , 𝛽
Dealing with Changes of Time-Aware Processes | Andreas Lanz, Manfred Reichert |
CSTN Transformation*
Activity
A 𝑥, 𝑦 AS AE
0,∞ , 𝛽
𝑥,𝑦 , 𝛽
branch 𝑃
XOR-Split
0,∞ , 𝑃𝛽
observation node
[0,∞],𝛽
0, 1 ,𝛽 XS XE P? P? X
true
false
0,∞ , ¬𝑃𝛽
branch ¬𝑃
Similar: XOR-Join, AND-Split, AND-Join
* [Lanz et al., 13]

10. Modeling Time-Aware Processes – Minimal Network
Dealing with Changes of Time-Aware Processes | Andreas Lanz, Manfred Reichert |
Modeling Time-Aware Processes – Minimal Network
Process Model
Minimal Time Model
CSTN Time Model
Restricted constraints
Derived interdependencies
Constraint between any pair of time points

11. Modeling Time-Aware Processes – Temporal Consistency
Dealing with Changes of Time-Aware Processes | Andreas Lanz, Manfred Reichert |
Modeling Time-Aware Processes – Temporal Consistency
 it exists a minimal time model
Weak consistency of the CSTN can be check using the algorithm proposed by [Tsamardinos et al., 03]
Definition (Temporal Consistency).
A time-aware process model is denoted as temporally consistent iff the corresponding time model (i.e., its CSTN representation) is (weakly) consistent.

12. Executing Time-Aware Processes
Dealing with Changes of Time-Aware Processes | Andreas Lanz, Manfred Reichert |
Executing Time-Aware Processes
Time constraints need to be monitored during execution
Not sufficient to check consistency only at design time
Hidden interdependencies between time constraints may exist
May restrict explicit constraints
Need to be known when executing a process instance

13. Executing Time-Aware Processes
Dealing with Changes of Time-Aware Processes | Andreas Lanz, Manfred Reichert |
Executing Time-Aware Processes
Creation of a new process instance
Create a copy of the minimal time model
Update instance time model to current state
Updated activity execution times
Execute process instance activities

14. Change Operations for Time-Aware Processes
Dealing with Changes of Time-Aware Processes | Andreas Lanz, Manfred Reichert |
Motivation
Time-Aware Processes
Change Operations for Time-Aware Processes
Analyzing the Effects
Conclusion

15. Change Operations – Basic Procedure
Dealing with Changes of Time-Aware Processes | Andreas Lanz, Manfred Reichert |
Change Operations – Basic Procedure
Check applicability of change operation
Update process model
Update instance time model
Restore minimality of instance time model
Update hidden interdependencies

16. Insert Activity Serially
Dealing with Changes of Time-Aware Processes | Andreas Lanz, Manfred Reichert |
Insert Activity Serially
𝒄 𝒎𝒂𝒙 ≥ 𝒅 𝒎𝒊𝒏
< 𝑚𝑎𝑥 𝑐 𝑚𝑖𝑛 , 𝑑 𝑚𝑖𝑛 , 𝑐 𝑚𝑎𝑥 ,𝛽>

17. Insert Activity Conditionally
Dealing with Changes of Time-Aware Processes | Andreas Lanz, Manfred Reichert |
Insert Activity Conditionally
𝒄 𝒎𝒂𝒙 ≥ 𝒅 𝒎𝒊𝒏

18. Insert Time Lag 𝒄 𝒎𝒂𝒙 ≥ 𝒕 𝒎𝒊𝒏 𝒄 𝒎𝒊𝒏 ≤ 𝒕 𝒎𝒂𝒙
Dealing with Changes of Time-Aware Processes | Andreas Lanz, Manfred Reichert |
Insert Time Lag
𝒄 𝒎𝒂𝒙 ≥ 𝒕 𝒎𝒊𝒏
𝒄 𝒎𝒊𝒏 ≤ 𝒕 𝒎𝒂𝒙

19. Dealing with Changes of Time-Aware Processes | Andreas Lanz, Manfred Reichert | 11-09-2014
Change Operations

20. Analyzing the Effects Motivation Time-Aware Processes
Dealing with Changes of Time-Aware Processes | Andreas Lanz, Manfred Reichert |
Motivation
Time-Aware Processes
Change Operations for Time-Aware Processes
Analyzing the Effects
Conclusion

21. Applying multiple change operations
Dealing with Changes of Time-Aware Processes | Andreas Lanz, Manfred Reichert |
Applying multiple change operations
Minimality of the instance time model needs to be restored after each change operation
Time complexity 𝑂 𝑛 3 2 𝑚
𝑛 number of time points
𝑚 number of observation time points
Significant delays may occur
Possible Solution: Approximation
How can the “new” instance time model be approximated?
Restore minimality of instance time model
Update hidden interdependencies
Approximate new instance time model
Update instance time model
Update process model
Check applicability of change operation
Restore minimality of instance time model
Update hidden interdependencies
Update instance time model
Update process model
Check applicability of change operation

22. Analyzing the Effects Proof in [Lanz et al., 14a]
Dealing with Changes of Time-Aware Processes | Andreas Lanz, Manfred Reichert |
Analyzing the Effects
Proof in [Lanz et al., 14a]

23. Example Theorem 1 – Insert Serial
Dealing with Changes of Time-Aware Processes | Andreas Lanz, Manfred Reichert |
Example Theorem 1 – Insert Serial
𝜹= 𝒅 𝒎𝒊𝒏 − 𝒄 𝒎𝒊𝒏
𝑎 𝑚𝑖𝑛 ≤ 𝒂 ′ 𝒎𝒊𝒏 ≤𝑎 𝑚𝑖𝑛 +𝜹
𝑎 𝑚𝑎𝑥 ≥ 𝒂 ′ 𝒎𝒂𝒙 ≥𝑎 𝑚𝑎𝑥 −𝜹
𝑏 𝑚𝑖𝑛 ≤ 𝒃′ 𝒎𝒊𝒏 ≤𝑏 𝑚𝑖𝑛 +𝜹
𝑏 𝑚𝑎𝑥 ≥ 𝒃 ′ 𝒎𝒂𝒙 ≥𝑏 𝑚𝑎𝑥 −𝜹

24. Multiple Change Operations – Optimized Procedure
Dealing with Changes of Time-Aware Processes | Andreas Lanz, Manfred Reichert |
Multiple Change Operations – Optimized Procedure
Restore minimality of instance time model
Update hidden interdependencies
Check applicability of change operation
Update process model
Update instance time model
Approximate new instance time model
Restore minimality of instance time model
Update hidden interdependencies

25. Applying Multiple Change Operations
Dealing with Changes of Time-Aware Processes | Andreas Lanz, Manfred Reichert |
Applying Multiple Change Operations
4≤14 − 𝛿
4≤10
4≤7
5≤4
5≤10 − 𝛿
5≤6
restore minimality

26. Applying Multiple Change Operations
Dealing with Changes of Time-Aware Processes | Andreas Lanz, Manfred Reichert |
Applying Multiple Change Operations
Approximate the resulting temporal properties of the entire process instance
Significantly reduces the complexity of applying multiple change operations
Actual saving depend on the “strictness” of the temporal constraints and the changes to be performed

27. Conclusion Motivation Time-Aware Processes
Dealing with Changes of Time-Aware Processes | Andreas Lanz, Manfred Reichert |
Motivation
Time-Aware Processes
Change Operations for Time-Aware Processes
Analyzing the Effects
Conclusion

28. Summary Time is a fundamental concept for business processes
Dealing with Changes of Time-Aware Processes | Andreas Lanz, Manfred Reichert |
Summary
Time is a fundamental concept for business processes
Flexibility is important for time-aware processes
Change operations for time-aware processes instances
Approximation for applying multiple changes
Significantly reduces the complexity of applying multiple change operations
Proof-of-Concept implementation as part of the ATAPIS Toolset

29. Ongoing and Future Work
Dealing with Changes of Time-Aware Processes | Andreas Lanz, Manfred Reichert |
Ongoing and Future Work
Investigating more complex change patterns
Application of presented results to process evolution
Integration of advanced time-management capabilities into the ATAPIS Toolset
Questions?
dbis.info/atapis

30. Dealing with Changes of Time-Aware Processes | Andreas Lanz, Manfred Reichert | 11-09-2014
References
[Hunsberger et al., 12] Luke Hunsberger, Roberto Posenato, and Carlo Combi. The dynamic controllability of conditional STNs with uncertainty. In Proceedings of the Planning and Plan Execution for Real-World Systems: Principles and Practices (PlanEx), [Lanz et al., 13] Andreas Lanz, Roberto Posenato, Carlo Combi, and Manfred Reichert. Controllability of time-aware processes at run time. In Proceedings of the 21st International Conference on Cooperative Information Systems (CoopIS'13), pages Springer, [Lanz et al., 14] Andreas Lanz, Barbara Weber, and Manfred Reichert. Time patterns for process- aware information systems. Requirements Engineering, 19(2): , [Lanz et al., 14a] Andreas Lanz and Manfred Reichert. Process change operations for time-aware processes. Technical Report UIB , University of Ulm, [Tsamardinos et al., 03] Ioannis Tsamardinos, Thierry Vidal, and Martha E. Pollack. CTP: A new constraint-based formalism for conditional, temporal planning. Constraints, 8(4): , 2003.