Talk at the First International Workshop on Collaborative Modelling in MDE (COMMitMDE), St Malo, France, 2016.
Abstract:
Due to the increase of their complexity, currently engineered systems cannot be developed by one individual, but are a product of a collaboration between multiple stakeholders who develop the system from different domain-specific views. Inconsistencies between views, however, hinder collaboration and therefore, must be managed. Since the encountered inconsistencies may potentially disappear as the natural consequence of a design workflow, tolerating them to a given extent may be desirable and can lead to a more efficient collaboration. A key to reason about tolerance is the quantification of the impact of single inconsistencies on the overall system design.
In this paper we present a quantification model for semantic inconsistencies in discrete and continuous systems. We investigate characteristic behavioral patterns of inconsistencies based on this model and identify the links with various forms of tolerance. Finally, we discuss the directions of further expanding the approach required for a comprehensive inconsistency tolerance technique.
Towards Inconsistency Tolerance by Quantification of Semantic Inconsistencies
1. Towards Inconsistency Tolerance
by Quantification of Semantic Inconsistencies
Istvan David, Eugene Syriani, Clark Verbrugge, Didier Buchs,
Dominique Blouin, Antonio Cicchetti, Ken Vanherpen
Saint-Malo, 04.10.2016.
3. Resolving inconsistencies
Automated Guided Vehicle (AGV)
INCONSISTENCY
An inconsistency is present if two or more statements are made that
are not jointly satisfiable [such as a] failure of an equivalence test,
non-conformance to a standard or constraint and the violation of
physical or mathematical principles (Herzig)
4. Resolving inconsistencies
Automated Guided Vehicle (AGV)
INCONSISTENCY
Let’s stop and
resolve!
An inconsistency is present if two or more statements are made that
are not jointly satisfiable [such as a] failure of an equivalence test,
non-conformance to a standard or constraint and the violation of
physical or mathematical principles (Herzig)
5. Resolving inconsistencies
Automated Guided Vehicle (AGV)
INCONSISTENCY
Let’s stop and
resolve!
An inconsistency is present if two or more statements are made that
are not jointly satisfiable [such as a] failure of an equivalence test,
non-conformance to a standard or constraint and the violation of
physical or mathematical principles (Herzig)
Increase
platform size
Increase
battery size
6. Resolving inconsistencies
Automated Guided Vehicle (AGV)
INCONSISTENCY
Let’s stop and
resolve!
An inconsistency is present if two or more statements are made that
are not jointly satisfiable [such as a] failure of an equivalence test,
non-conformance to a standard or constraint and the violation of
physical or mathematical principles (Herzig)
Increase
platform size
Increase
battery size
7. Resolving inconsistencies
Automated Guided Vehicle (AGV)
INCONSISTENCY
Let’s stop and
resolve!
An inconsistency is present if two or more statements are made that
are not jointly satisfiable [such as a] failure of an equivalence test,
non-conformance to a standard or constraint and the violation of
physical or mathematical principles (Herzig)
Increase
platform size
Increase
battery size
Semantic inconsistencies
8. Resolving inconsistencies
Automated Guided Vehicle (AGV)
INCONSISTENCY
An inconsistency is present if two or more statements are made that
are not jointly satisfiable [such as a] failure of an equivalence test,
non-conformance to a standard or constraint and the violation of
physical or mathematical principles (Herzig)
Let’s stop and
resolve wait?
Increase
platform size
Increase
battery size
Semantic inconsistencies
9. Resolving inconsistencies
Automated Guided Vehicle (AGV)
INCONSISTENCY
An inconsistency is present if two or more statements are made that
are not jointly satisfiable [such as a] failure of an equivalence test,
non-conformance to a standard or constraint and the violation of
physical or mathematical principles (Herzig)
Let’s stop and
resolve wait?
Increase
platform size
Increase
battery size
Semantic inconsistencies
10. Resolving inconsistencies
Automated Guided Vehicle (AGV)
INCONSISTENCY
An inconsistency is present if two or more statements are made that
are not jointly satisfiable [such as a] failure of an equivalence test,
non-conformance to a standard or constraint and the violation of
physical or mathematical principles (Herzig)
Let’s stop and
resolve wait?
Increase
platform size
Increase
battery size
Semantic inconsistencies
11. Resolving inconsistencies
Automated Guided Vehicle (AGV)
INCONSISTENCY
An inconsistency is present if two or more statements are made that
are not jointly satisfiable [such as a] failure of an equivalence test,
non-conformance to a standard or constraint and the violation of
physical or mathematical principles (Herzig)
Let’s stop and
resolve wait?
Increase
platform size
Increase
battery size
Semantic inconsistencies
12. Increase
platform size
Increase
battery size
Resolving inconsistencies
Automated Guided Vehicle (AGV)
INCONSISTENCY
An inconsistency is present if two or more statements are made that
are not jointly satisfiable [such as a] failure of an equivalence test,
non-conformance to a standard or constraint and the violation of
physical or mathematical principles (Herzig)
Let’s stop and
resolve wait?
Semantic inconsistencies
When to resolve an inconsistency?
13. Increase
platform size
Increase
battery size
Resolving inconsistencies
Automated Guided Vehicle (AGV)
INCONSISTENCY
An inconsistency is present if two or more statements are made that
are not jointly satisfiable [such as a] failure of an equivalence test,
non-conformance to a standard or constraint and the violation of
physical or mathematical principles (Herzig)
Let’s stop and
resolve wait?
Semantic inconsistencies
When to resolve an inconsistency?
14. Increase
platform size
Increase
battery size
Resolving inconsistencies
Automated Guided Vehicle (AGV)
INCONSISTENCY
An inconsistency is present if two or more statements are made that
are not jointly satisfiable [such as a] failure of an equivalence test,
non-conformance to a standard or constraint and the violation of
physical or mathematical principles (Herzig)
Let’s stop and
resolve wait?
Semantic inconsistencies
When to resolve an inconsistency?
Always consistent
Eventually consistent
±10%
15. When to resolve an inconsistency?
Q1: How severe the inconsistency of the whole model space is?
Q2: What are the chances that the inconsistency gets resolved
without intervening?
16. When to resolve an inconsistency?
Q1: How severe the inconsistency of the whole model space is?
Q2: What are the chances that the inconsistency gets resolved
without intervening?
17. When to resolve an inconsistency?
Q1: How severe the inconsistency of the whole model space is?
Q2: What are the chances that the inconsistency gets resolved
without intervening?
Number of detected inconsistency instances,
number of object affected, severity and
priority (labels), etc
18. When to resolve an inconsistency?
Divergent viewpoints (Easterbrook et al)
Pre- and post-conditions (FOL)
Manual triggering of resolution
Q1: How severe the inconsistency of the whole model space is?
Q2: What are the chances that the inconsistency gets resolved
without intervening?
Number of detected inconsistency instances,
number of object affected, severity and
priority (labels), etc
19. When to resolve an inconsistency?
Divergent viewpoints (Easterbrook et al)
Pre- and post-conditions (FOL)
Manual triggering of resolution
Q1: How severe the inconsistency of the whole model space is?
Q2: What are the chances that the inconsistency gets resolved
without intervening?
Number of detected inconsistency instances,
number of object affected, severity and
priority (labels), etc
20. When to resolve an inconsistency?
Divergent viewpoints (Easterbrook et al)
Pre- and post-conditions (FOL)
Manual triggering of resolution
Q1: How severe the inconsistency of the whole model space is?
Q2: What are the chances that the inconsistency gets resolved
without intervening?
Number of detected inconsistency instances,
number of object affected, severity and
priority (labels), etc
We give formal foundations for quantifying how
severe semantic inconsistencies are in multi-view
collaborative settings.
21. Quantifying inconsistencies
• Models expressed with operational semantics, consisting of traces on
given states
• Calculate the distance of the traces
22. Quantifying inconsistencies
• Models expressed with operational semantics, consisting of traces on
given states
• Calculate the distance of the traces
23. Quantifying inconsistencies
• Models expressed with operational semantics, consisting of traces on
given states
• Calculate the distance of the traces
Mechanical view Electrical view
Battery support size Battery size
24. Quantifying inconsistencies
• Models expressed with operational semantics, consisting of traces on
given states
• Calculate the distance of the traces
Mechanical view Electrical view
Battery support size Battery size
=
25. Quantifying inconsistencies
• Models expressed with operational semantics, consisting of traces on
given states
• Calculate the distance of the traces
Mechanical view Electrical view
Battery support size Battery size
=
26. Quantifying inconsistencies
• Models expressed with operational semantics, consisting of traces on
given states
• Calculate the distance of the traces
Mechanical view Electrical view
Battery support size Battery size
=
Consistency: δ=0
27. Quantifying inconsistencies
• Models expressed with operational semantics, consisting of traces on
given states
• Calculate the distance of the traces
Mechanical view Electrical view
Battery support size Battery size
=
Consistency: δ=0
• Allow accumulation over time
28. Quantifying inconsistencies
• Models expressed with operational semantics, consisting of traces on
given states
• Calculate the distance of the traces
Mechanical view Electrical view
Battery support size Battery size
=
Consistency: δ=0
• Allow accumulation over time
Now we can assess the severity of inconsistencies in
terms of a formalized view distance metric
35. Future work
• Apply the techniques in a process modeling and execution tool
• Design-time/run-time trade-off
• Automation of tolerance rule specification
• Predictive impact analysis
• Modeling engineering operations
• Resolution scheduling
• Proving global (in)consistency