Research @ RELEASeD (presented at SATTOSE2013)

Research @ RELEASeD
The RELEASeD Laboratory
Achieving Excellence in Software Development
released.info.ucl.ac.be

Monday 8 July 2013

1

RELEASeD – the lab
REsearch Laboratory on
software Evolution And Software Development
Computing Science Engineering Department (INGI)
Institute of Information and Communication Technologies,
Electronics and Applied Mathematics (ICTEAM)
Université catholique de Louvain (UCL)
Monday 8 July 2013

2

RELEASeD – the research
Current research domains
Software Evolution Technology
Software Development Technology
Context-Oriented Programming
Strong focus on tools, techniques and languages
Monday 8 July 2013

3

RELEASeD – the team
Kim Mens , Professor, Head of RELEASeD group
Kim.Mens@uclouvain.be
Sergio Castro , Teaching and Research Assistant
Sergio.Castro@uclouvain.be
Ángela Lozano , Post-Doctoral Researcher
Angela.Lozano@uclouvain.be
Nicolás Cardozo , PhD student
nicolas.cardozo@uclouvain.be
Sebastián González , Post-Doctoral Researcher
s.gonzalez@uclouvain.be
Monday 8 July 2013

4

Some focused presentations
Context-Oriented Programming [Sebastian, Nicolas, Kim]
Recent research on COP : Version Contexts , Context Traits &
Context Petri Nets

Software Quality Assurance [Angela, Kim]
Tools and techniques to improve the structural quality of
software : Usage Contracts & Mendel

Java-Prolog Interoperability [Sergio, Kim]
Monday 8 July 2013

5

SQA @ RELEASeD
Kim Mens, Angela Lozano

Monday 8 July 2013

6

Software Quality Assurance
Mendel
a code recommendation tool that provides structural
recommendations to software developers based on local
structural properties in the source code of a system
Usage Contracts
a unit testing-like approach to enable developers to
express and verify structural coding conventions in a
seamless way
Monday 8 July 2013

7

COP @ RELEASeD
Kim Mens, Sebastian Gonzalez, Nicolas Cardozo

Monday 8 July 2013

8

COP @ RELEASeD
Research on COP since 10 years
Ambience [2008], Subjective-C [2010], SCOPJ [2011],
Phenomenal Gem [2012]
Most recent developments
Version Contexts [UCL 2013]
Context Traits [AOSD 2013]
Context Petri Nets [UCL 2013]
Monday 8 July 2013

9

Motivation : Towards a Mindset Shift
From:

To:

Programming in Isolation

Programming with Context

?

?
?

forward!
?

?
?

Monday 8 July 2013

11

From:

To:

Programming in Isolation

Programming with Context

Contexts
Situations of the surrounding execution
environment to which speciﬁc behavior
is associated

Monday 8 July 2013

12

?

?
?

forward!
?

?
?

Monday 8 July 2013

Caution
Design patterns
Plugin architectures
Conditional statements
...
13

Conditional Statements : some variant of ...
m(a) {
!

if (

) { IE logic }

else if (

) { Opera logic }

else if (

) { Chrome logic }

else if (

) { Safari logic }

else if (

) { Firefox logic }

else

{ default logic }

}
Monday 8 July 2013

14

Conditional Statements : some variant of ...
m(a) {
!

if (

) { IE logic }

else if (

) { Opera logic }

else if (

) { Chrome logic }

else if (

) { Safari logic }
Tangled

else if (

Adaptable

else

) { Firefox logic
{ default logic

Scattered
}
Fixed
No reuse
}
Complex logic

}
Monday 8 July 2013

14

With design patterns ...
m(a) {
strategy
strategy.m(a)
}

m(a) { IE strategy }
m(a) { Opera strategy }
m(a) { Chrome strategy }
m(a) { Safari strategy }
m(a) { Firefox strategy }
m(a) { default strategy }

Modular
Open
Monday 8 July 2013

Infrastructural burden
Anticipated adaptation points
15

A Toy Example
(really)
(default context)
adaptation
CODE: move toward
pacman

CODE: move randomly
in the maze

Context Expert

Monday 8 July 2013

adaptation

Context Beginner

17

An excerpt of the COP implementation of Pac-Man™ in
Phenomenal Gem
!"#$$!"#$%&
Default behaviour
!!%&'!'()*&+,($%-&-$.
!!!!!"#$%&'"()*%"+,-.)(/0"1-"23%"(,4%
!!!!!"5
!!&(%
&(%

/#$%&0'()*&+,($%-&-$.!!"()*%"+,-.)(/0

Monday 8 July 2013

18

Phenomenal Gem
!"#$$!"#$%&
!!%&'!'()*&+,($%-&-$.
!!!!!"#$%&'"()*%"+,-.)(/0"1-"23%"(,4%
!!!!!"5
!!&(%
&(%
!
+1(+2&,3$.&+1&!4!5$.&+1&0.+6
+1(+2&,3$.&+1&0*)),*)*(&*&-$.)"#$%&7!*+,%#-&.,/$0-0/(7!-1+&7!
22,1/!!3!!"#$%&'"()*%"2)6,+."$,&(,-24
45
/#$%&0'()*&+,($%-&-$.!!"()*%"+,-.)(/0
+1(+2&,3$.&+1&0*3&-8*&+
/#$%&0'()*&+,($%-&-$.2!"()*%"2)6,+."$,&(,Monday 8 July 2013

18

Phenomenal Gem
!"#$$!"#$%&
!!%&'!'()*&+,($%-&-$.
!!!!!"#$%&'"()*%"+,-.)(/0"1-"23%"(,4%
!!!!!"5
Explicit
!!&(%
Contexts
&(%
Deﬁnition of Adaptations
!
(even dynamically)
+1(+2&,3$.&+1&!4!5$.&+1&0.+6
+1(+2&,3$.&+1&0*)),*)*(&*&-$.)"#$%&7!*+,%#-&.,/$0-0/(7!-1+&7!
22,1/!!3!!"#$%&'"()*%"2)6,+."$,&(,-24
45
(De)activation of contexts
/#$%&0'()*&+,($%-&-$.!!"()*%"+,-.)(/0
Dynamic Behaviour
+1(+2&,3$.&+1&0*3&-8*&+
Adaptation
/#$%&0'()*&+,($%-&-$.2!"()*%"2)6,+."$,&(,Monday 8 July 2013

18

Some COP language extensions we have been working on ...
Ambience [CLOS]
Subjective-C [Objective-C]
SCopJ [Java]
Phenomenal Gem [Ruby]
Context Traits [JavaScript]
plus a small educational prototype in Smalltalk
Monday 8 July 2013

19

Version Contexts
Etienne Martel © June 2013 (MSc thesis, UCLouvain)
Advisors: Sebastian Gonzalez, Kim Mens

Monday 8 July 2013

20

Motivation : Subversion
Maintaining backward compatibility can lead to software erosion
e.g., in the source code of Subversion original functions like
!"#$%&'(#)$%*(%+,-)./!"#$%&'(#)$%,00')./!"#$%&'(#)$122./333

were subsequently extended with new functionality and parameters
!"#$%&'(#)$%*(%+,-)4.//!"#$%&'(#)$%*(%+,-)5
!"#$%&'(#)$%,00')./!"#$%&'(#)$%,00')4./333./
!"#$%&'(#)$%,00')6
!"#$%&'(#)$122./!"#$%&'(#)$1224./333./!"#$%&'(#)$1227

Monday 8 July 2013

21

Motivation : Skype

Monday 8 July 2013

22

Motivation : Skype

Monday 8 July 2013

23

Motivation : Skype
Need for life software updates
without
Service interruption
program state loss
losing old program version
...
Monday 8 July 2013

23

The idea
Exploit COP by considering versions as a kind of context that
can be dynamically (de)activated
activating that context = moving to a new version
deactivating = reverting to an older version
Original use of COP in a way it hasn’t been speciﬁcally
conceived for
Monday 8 July 2013

24

Version Contexts
Version Contexts extend Phenomenal Gem
Engine

VersionContext 1
Monday 8 July 2013

Pacman

VersionContext 2
25

A Quick Demo

Monday 8 July 2013

26

Implementation Issues
Instance migration
migrating an existing instance to a new class
some capabilities are missing in Ruby
Instance state
Version mappings
Monday 8 July 2013

27

Instance Migration
Object behavior must match the class deﬁnition of the
currently active version context

TRADITIONAL

Monday 8 July 2013

NEW

NEW WITH 2 CLASSES

28

Instance State
Storing the values of attributes of each object for every
version is inefﬁcient.
Solution :
store them for only one of the versions, e.g. the latest
redirect attribute access for the other versions

Monday 8 July 2013

29

Version Mappings
But what if the representation of an attribute evolves?
For example, from money in $ to money in !
VERSION 1

VERSION 2

{ Entity@money in v1 is represented differently in v2 }
Monday 8 July 2013

30

Version Mappings
Or what if a class is renamed from one version to another?

VERSION 1

VERSION 2

{ Entity in v1 = Pacman in v2 }
Monday 8 July 2013

31

Version Mappings
Or what if an attribute gets renamed?

VERSION 1

VERSION 2

{ Entity@location in v1 = Pacman@position in v2 }
Monday 8 July 2013

32

Version Mappings
Need for version mappings to map from the old version to
the new one (and back)
(Bidirectional) functions deﬁned by the programmer to
render explicit these implicit changes between versions
class Pacman
attr_accessor :money
# 1 US$ = 27.435 FB
result_mapping :@money, lambda {|val| val * 27.435 },
lambda {|val| val / 27.435 }
# ...
end
Monday 8 July 2013

33

Version Mappings
Currently supported version mappings :
Name version mappings (classes, methods, attributes)
Parameter version mappings
Result version mappings
More version mappings could (should) be added ...
Monday 8 July 2013

34

Current Implementation Restrictions
Linear versioning vs branch versioning
Granularity of an update: full version vs delta version
No support for concurrency

Monday 8 July 2013

35

Context Traits
(very brieﬂy)
Sebastian Gonzalez, Marius Colacioiu, Kim Mens, Walter Cazzola
© AOSD 2013

Monday 8 July 2013

36

The idea
Dynamic Behaviour Adaptation Through Run-Time Trait
Recomposition
Paper presented at AOSD•Composition 2013
Adopt the notion of Traits, a static composition mechanism,
in a more dynamic setting, and use it as a building block for
implementing COP
Monday 8 July 2013

37

Context-Driven Trait Composition
T2
T1
T3

TD

T5

Monday 8 July 2013

T4

38

T2
T1
T3

TD
Default behaviour

T5

Monday 8 July 2013

T4

38

T2
T1
T3

TD

Context-speciﬁc adaptations

T5

Monday 8 July 2013

T4

38

T2

T1 T T3
D
T5

T4

Monday 8 July 2013

38

T2

T1 T T3
D
T4

T5

Monday 8 July 2013

38


T1TT2
D
T4

T3

T5

Monday 8 July 2013

38

Contributions
Traits are convenient units of adaptation
more than single methods, less then full classes
Support for ﬂexible composition policies
e.g., non-linear (default policy = by activation age)
Support for overriding behaviour with “proceed”
JavaScript allows for easy deﬁnition of contexts, traits and
composition
build on top op traits.js [Van Cutsem & Miller 2011]
Monday 8 July 2013

39

In the paper...
‣ Contexts
‣ Traits
‣ Context-Driven Trait Compositions
‣ Composition Policies
‣ Behaviour Extensibility
‣ Context Traits in JavaScript
‣ Implementation Notes
‣ Case Studies
‣ Related Work
‣ Future Work
Monday 8 July 2013

40

Context Petri Nets
Nicolas Cardozo © 2013
PhD thesis, advisors: Kim Mens & Theo D’Hondt

Monday 8 July 2013

41

Research question
How to increase conﬁdence in context-oriented systems that
can dynamically adapt their behavior as context are being
activated and deactivated ?
Support for context dependency relations to express
allowed and disallowed interactions between contexts
Plus a mechanism to ensure consistency of the system
with respect to those declared dependencies
both statically and dynamically
Monday 8 July 2013

42

Context Dependency Relations

Monday 8 July 2013

43

CompassGPSAntenna

Context
dependency
relations
Different types :
• Exclusion
• Causality
• Implication
• Requirement
• Conjunction
• (Disjunction)
• (Suggestion)

Positioning

Compass

GPSAntenna

Tour
ItineraryPOIOrder

FreeWalk

GuidedTour

HighBattery

UserLanguage |
Language

UserLanguage

CategoryPOIOrder

LocationLanguage

LowBattery

AudioStream

VideoStream

UserPreferences
TimeOfDay

3G

Connectivity

Wiﬁ

SimpleInterface

ColoredCategory

LowMemory

Night

Afternoon

Subjective-C
RefreshPOIMap
Monday 8 July 2013

Morning

ConnectivityLowMemory
43

Research goal
How to ensure that
the declared model makes sense
verifying coherence and correctness of interactions
the model is respected by the system at run-time
ensure run-time consistency
Monday 8 July 2013

44

Context Petri nets
0.1 slide 14

1

Singleton CoPN
0.1 slide 14
A context is represented by a context Petri net (CoPN)
deﬁned as C = Pc , Pt , Te , Ti , f, f◦ , ρ, L, m0 , Σ

Subjective-C
8%,#)(9):;,2(
2+:;$)+

92:;$)+

81%)'1)(:;,2(

Pc ∩ Pt = φ
Te ∩ Ti = φ
(Pc ∪ P ) ∩ (Te ∪ Ti ) = φ 2+:=;$)+
*3:;$)+ t
f : (P × T × L) ∪ (T × P × L) −→ Z∗
f◦ : P × T −→ {0, 1}
;$)+

0.2 slide 21

ρ : T −→ Z∗
∀ t ∈ Te , ρ(t) = 0
∀ t ∈ Ti , ρ(t) 0
m0 : P × L −→ Z∗

)+*3:;$)+

92:=;$)+

82(1%)'1)(:;,2(

◦({CQ , CN , CM }, {E, CQ , CN , S, CM , CQ })

Context Petri nets.
Subjective-C: Bringing Context to Mobile Platform Programming.
González, Sebastián and Cardozo, Nicolás and Mens, Kim and Cádiz, Alfredo
http://released.info.ucl.ac.be/Tools/Context-PetriNets
and Libbrecht, Jean-Cristophe and Goffaux, Julien. International Conference ◦({C , C }, {S, C , C })
Q M
M Q
on Software Language Engineering (SLE’10).
Modeling and Analyzing Self-adaptive Systems With Context Petri
Nets. Cardozo, Nicolás and González, Sebastián and Mens, Kim and
Ragnhild Van Der Straeten and D’Hondt,Theo. International Symposium on
Theoretical Aspects of Software Engineering (TASE ’13).
Monday 8 July 2013

context
place

external
transition

activation
type

temporary
place

internal
transition

arcs

union({CQ , CM })

extS ({CQ , CM }, {S, CM , CQ })

inhibitor
arcs

46

0.1 slide 14


E, CA , CB
C, CA , CB
I, CA , CB

1

1
1

Pc ∩ Pt = φ
Te ∩ Ti = φ
(Pc ∪ Pt ) ∩ (Te
f : (P × T × L
f◦ : P × T −→

∧, CA1 , . . . , CAn

E, CA,,,C B@/A#%%?@+01=(-A/*/E,-//'*/=,!B
E, CAACCBBC
Q, C C,
E,
A

Context dependency relations

B

Five types of context dependency relations:
0.2 slide
Exclusion (E), Causality (C), Implication (I),
Requirement (Q), conjunction (∧)

C,C, ,CC,BCB
C, CA , A B @/A#%%:#+$#0-;1/=*/C/-/*/DB 1
CA C
I, I, ,CC,BCB @/A#%%=,0!#-0/(1/=*/C/-/*/DB
I, CA , CB
CA A

, .◦ n
∧,TCA T,, . f,. fC,Aρ,n m0 , Σ
t , C, 1
eA
∧,∧,1CiA,.1.. .. ,, .CACAL,@/A#%%:/(B+(!-0/(C'*/AC/./333./C/BB
n
@

15

0.3 slide 16

.

E,
Q, C , C@/A#%%67!+$0/(8-9(*/;,2(/#(%*/=(?B
Q, CAA,,CCBBB
Q, CA , A B
C C
*3:;$)+
∗

ρ : T −→ Z
8%,#)(9):;,2(
∀ t ∈ Te , ρ(t) = 0
φ I, CA , CB
∀ t ∈ Ti , ρ(t) 2+:;$)+
0
× ,P t , Te ,. T. ,,f, Z◦ , ρ, L, m00 Σ × L −→ Z∗
× L) .−→ f ∗
m, : P
∧, CA1 , i CAn
c P

C

;$)+

C, CA , CB

92:=!;$)+

S ⊆ S
2+:=;$)+R ⊆ R)+*3:;$)+
◦(S

92:;$)+

t , Te , Ti , f, f◦ , ρ, L, m0 , Σ
t , Te , Ti ,,f, f◦ , ρ, L, m0 , Σ
Q, C C
A

B

ρ : T −→ Z∗
Subjective-C
92:92+?
∀ t ∈ Te , ρ(t) = 0
∗
∗
ρ :: T −→ Z
ρt∈T , Z
)=φ
∀ T −→
8%,#)(9):=(? ρ(t) 0
∀ t ∈ P ie ,, L −→= ∗0
∗
∀ t ∈ T × ρ(t)2+:92+?
T × P × L) −→ Z
m0 : Te ρ(t) = 0
Z
*3:92+?
φ Te , Ti , f, f◦ , ρ, L, m0 , Σt ∈ Tii,, ρ(t) 0
∀ t ∈ T ρ(t) 0
φ
∀
},
t

)+*3:92+?

92+?

2+:=92+? slide 22
0.492:=92+?

[Subjective-C:
to Mobile
∗
P × L) −→Bringing Contextm0 :: P Platform Programming. SLE’10]
Z∗
× L −→ Z∗
P × L) −→analyzing self-adaptive systems with context Petri nets.TASE’13]
Z
m0 P × L −→ Z∗
[Modeling and

= Ti 8∪ {deac(A)}
Monday July 2013

ρ : T −→ Z∗

47

0.1 slide 14


E, CA , CB
C, CA , CB
I, CA , CB

1

1
1

Pc ∩ Pt = φ
Te ∩ Ti = φ
(Pc ∪ Pt ) ∩ (Te
f : (P × T × L
f◦ : P × T −→

∧, CA1 , . . . , CAn

E, CA,,,C B@/A#%%?@+01=(-A/*/E,-//'*/=,!B
E, CAACCBBC
Q, C C,
E,
A


B

Five types of context dependency relations:
0.2 slide
Exclusion (E), Causality (C), Implication (I),
Requirement (Q), conjunction (∧)

C,C, ,CC,BCB
C, CA , A B @/A#%%:#+$#0-;1/=*/C/-/*/DB 1
CA C
I, I, ,CC,BCB @/A#%%=,0!#-0/(1/=*/C/-/*/DB
I, CA , CB
CA A

, .◦ n
∧,TCA T,, . f,. fC,Aρ,n m0 , Σ
t , C, 1
eA
∧,∧,1CiA,.1.. .. ,, .CACAL,@/A#%%:/(B+(!-0/(C'*/AC/./333./C/BB
n
@

C, CA , CB

*3:A$'2

A$'2

92:=A$'2

S ⊆ S
2+:=A$'2R ⊆ R)+*3:A$'2
◦(S

92:A$'2

t , Te , Ti , f, f◦ , ρ, L, m0 , Σ
t , Te , Ti ,,f, f◦ , ρ, L, m0 , Σ
Q, C C
A

15

0.3 slide 16

.

E,
Q, C , C@/A#%%67!+$0/(8-9(*/C/#(%*/DB
Q, CAA,,CCBBB
Q, CA , A B
C C
∗

ρ : T −→ Z
∀ t ∈ Te , ρ(t) = 0
φ I, CA , CB
∀ t ∈ Ti , ρ(t) 2+:A$'2
0
× ,P t , Te ,. T. ,,f, Z◦ , ρ, L, m00 Σ × L −→ Z∗
× L) .−→ f ∗
m, : P
∧, CA1 , i CAn
c P

C

%#!)A/+15

B

ρ : T −→ Z∗
92:9$%
∀ t ∈ Te , ρ(t) = 0
∗
∗
ρ :: T −→ Z
ρ t ∈ T , ρ(t) 0
)=φ
∀ T −→ Z
∀ t ∈ P ie ,, L −→= ∗0
∗
∀ t ∈ T × ρ(t) 2+:9$%
T × P × L) −→ Z
m0 : Te ρ(t) = 0
Z
φ
∀ t ∈ T ρ(t) 0
φ
∀
}, Te , Ti , f, f◦ , ρ, L, m0 , Σt ∈ Tii,, ρ(t) 0
t

)+*3:9$%

*3:9$%

9$%

2+:=9$%
0.4

slide 22

92:=9$%

[Subjective-C:
to Mobile
∗
P × L) −→Bringing Contextm0 :: P Platform Programming. SLE’10]
Z∗
× L −→ Z∗
P × L) −→analyzing self-adaptive systems with context Petri nets.TASE’13]
Z
m0 P × L −→ Z∗
[Modeling and

= Ti 8∪ {deac(A)}
Monday July 2013

ρ : T −→ Z∗

48

C = Pc , ∈ tT Teρ(t),◦ 0◦ ,× T −→, Σ 1} ∗
, , Ti : P ρ, L, m0 {0,
∪ Ti ) = φ
∀ t P : i(P ,×f f, f ∪ (T × P × L) −→ Z
f
T × L)
∗C = P , P , T , T , f, f , ρ, L, m , Σ
) ∪ (T × P × L) −→ Z
m f◦ e P × T ◦−→∗ 0
c 0t : P: ×i L −→ Z {0, 1}
{0, 1}
Pc ∩ Pt = φ
ρ : T −→ Z∗

m0 : P × L −→ Z∗

Composition of CoPNs

P∩ P =
ρ : T −→ ∈ ∗Te , ρ(t) = 0
Te c ∩Tit = φ
∀t Z
Te ∩ P ) φ
0.2
(Pc ∪ Tit= ∩ (Te ∪ Ti ) = φ slide 16 ∀ t ∈ Te , ∈ Ti= 0
∀ t ρ(t) , ρ(t) 0
(Pc ∪ Pt ) ∩ (Te0.2) = φ
∪ T slide 16
∀ t ∈ T , ρ(t) 0
f f: :(P × T × L) ∪∪i (T × × L) L) −→ Z∗ m : P i× L P ×Z∗ −→ Z∗
× L) (T × P P × −→ Z∗
m0 : −→ L
(P × T
0
92:=92+?
92:92+?
f◦ ◦: :P × T −→ {0, 1}
92+?
)+*3:92+?
f P × T −→ {0, 1}
◦ :*3:92+? × ℘(R) → P × ℘(R) → P
℘(S ) ◦ : ℘(S )
Composition
◦ : ℘(S ) of contexts
S
Given a set × ℘(R) → Pand Sset S
S ⊆ a ⊆ 2+:92+?
2+:=92+?
of context dependency relationsR ⊆ R
R ⊆ R =(?((#%(!
cons(ext(union(S), R), R)
0.2 slide 16
the slide 16
0.2 composition is deﬁned as :


92:;$)+

92:=;$)+

◦ : ℘(S ) × ℘(R) → P

S

: ℘(S ) ×
0.3 ◦slide 22 ℘(R) → P
contexts
- S : ⊆ S =(?((#%(!./;,2(.
2+:;$)+
F((G1+
⊆
⊆RS R

0.3 slide 22

*3:;$)+

;$)+

2+:=;$)+

)+*3:;$)+

◦({CU , CL }, {∨, CU , CL })
;,2(

Exclusion
-R : ⊆ R (B) and
◦({CU , CL }, {∨, CU , CL })
◦({CU , (5)
causality CL }, {∨, CU , CL }) cons(ext(union(S), R), R)
92:=;$)+
92:;$)+
union({CU , CL })
relations
0.3 slide 22

union({CU , CL })

0.3 slide 22

union({CU , CL })
2+:=;$)+

2+:;$)+

◦({CU , CL }, {∨, CU , CL })
;$)+
ext∨*3:;$)+ L }, {∨, CU , CL })
({CU , C
;,2(

)+*3:;$)+

ext∨ ({CU , CL }, {∨, CU , CL }) ◦({CU , CL }, {∨, CU , CL })Pc = Pc ∪ {U|L}
union({C , C })

Pt

Pc = Pc ∪ {U|L}

U

ext∨ ({CU , CL }, {∨,92:=C2++L })
CU , C

L 92:C2++

Pt = Pt ∪ {P r(U|L), P r(¬U|L)}

Pc = Pc ∪ {U|L}

ext∨ ({Cunion({CU , CCL}) Ti ∪ {act(U|L), deac(U|Ln )}
U , CL }, {∨, CU , T })
2+:C2++
2+:=C2++
L =

= Pt ∪ {P r(U|L), P r(¬U|L)}

TMonday 8 July∪ {act(U|L), deac(U|Ln )}
i = Ti 2013

i

Pc

*3:C2++

)+*3:C2++

Pt = Pt ∪ {P r(U|L), P r(¬U|L)}
= PcF((G1+
∪ {U|L}
ρ(t) if t inTe ∪ Ti

49
ρ (t) =
C2++

C = Pc , ∈ tT Teρ(t),◦ 0◦ ,× T −→, Σ 1} ∗
, , Ti : P ρ, L, m0 {0,
∪ Ti ) = φ
∀ t P : i(P ,×f f, f ∪ (T × P × L) −→ Z
f
T × L)
∗C = P , P , T , T , f, f , ρ, L, m , Σ
) ∪ (T × P × L) −→ Z
m f◦ e P × T ◦−→∗ 0
c 0t : P: ×i L −→ Z {0, 1}
{0, 1}
Pc ∩ Pt = φ
ρ : T −→ Z∗


m0 : P × L −→ Z∗

P∩ P =
ρ : T −→ ∈ ∗Te , ρ(t) = 0
Te c ∩Tit = φ
∀t Z
T ∩ Ti = φ
∀ t ∈ Te , ∈ T = 0
∀ t ρ(t) i , ρ(t) 0
if A ∈ t •(PcA∪ Pt ) ∧ (T= ∪ Ti )l = φ slide 16
∧e ∈ •t ∩ p e B ∧ 0.2
/
∈L
(Pc ∪ Pt ) ∩ (Te0.2) = φ
∪ T slide 16
∀ t ∈ T , ρ(t) 0
× L) (T × P P × −→ Z∗
m0 : −→ L
otherwise (P × T
0
:/DE 92:92+?
92:=92+?
f◦ ◦: :P × T −→ {0, 1}
92+?
)+*3:92+?
f P × T −→ {0, 1}
◦ :*3:92+? × ℘(R) → P × ℘(R) → P
℘(S ) ◦ : ℘(S )
Composition
if A ∈ ℘(S∧ A ∈ •t ∧ → = B ∧ l ∈ L
◦ : ta•set ×/contextsPand a set S
) of ℘(R) p
S
Given
S ⊆ S ⊆ 2+:92+?
2+:=92+?
otherwise dependency relationsR
of context
R ⊆ R ⊆ R

0.2 slide 16
the slide 16


◦ : ℘(S ) × ℘(R) → P

: ℘(S ) ×
0.3 ◦slide 22 ℘(R) → P
contexts
-) :⊆ S → P
℘(S S × ℘(R) =(?((#%(!./;,2(.
S

F((G1+
⊆
⊆RS R

0.3 slide 22
92:;$)+

◦({CU , CL }, {∨, CU , CL })

92:=;$)+

Exclusion
-R : ⊆ R (B) and
◦({CU , CL }, {∨, CU , CL })
2+:=;$)+
2+:;$)+
◦({CU , (5)
)+*3:;$)+
*3:;$)+
;$)+
ons(ext(union(S), R), R)
union({CU , CL })
relations
0.3 slide 22

union(S)
union({CU , CL })

union({CU , CL })

0.3 slide 22

◦({CU , CL }, {∨, CU , CL })
ext∨ ({CU , CL }, {∨, CU , CL })

ext(S, R)

ext∨ ({CU , CL }, {∨, CU , CL }) ◦({CU , CL }, {∨, CU , CL })Pc = Pc ∪ {U|L}
union({CU , CL })
ext∨ ({CU , CL }, {∨,92:=C2++L })
CU , C
cons(S, R)
92:C2++

Pt = Pt ∪ {P r(U|L), P r(¬U|L)}
Pc = Pc ∪ {U|L}

Q, CH , CV
Pc = Pc ∪ {U|L}

2+:C2++
2+:=C2++
L })
Pt = Pt ∪ {P r(U|L), P r(¬U|L)} U , CL }, {∨, CU , Ti = *3:C2++
C2++
)+*3:C2++

C, CC , CV

Pt = Pt ∪ {P r(U|L), P r(¬U|L)}
Pc = Pc ∪ {U|L}


i = Ti 2013
49

ρ (t) =

C = Pc , ∈ tT Teρ(t),◦ 0◦ ,× T −→, Σ 1} ∗
, , Ti : P ρ, L, m0 {0,
∪ Ti ) = φ
∀ t P : i(P ,×f f, f ∪ (T × P × L) −→ Z
f
T × L)
∗C = P , P , T , T , f, f , ρ, L, m , Σ
) ∪ (T × P × L) −→ Z
m f◦ e P × T ◦−→∗ 0
c 0t : P: ×i L −→ Z {0, 1}
{0, 1}
Pc ∩ Pt = φ
ρ : T −→ Z∗


m0 : P × L −→ Z∗

P∩ P =
ρ : T −→ ∈ ∗Te , ρ(t) = 0
Te c ∩Tit = φ
∀t Z
T ∩ Ti = ∩
φ
∀ t ∈ Te , ∈ T = 0
∀ t ρ(t) i , ρ(t) 0
if A ∈ t •(PcA∪/Pt ) ∧ (T= ∪ Ti )l = L slide 16
∧eA ∈ •t ∧ p = B ∧ 0.2
/ •t p e B ∧ l ∈ φ
∈L
if A ∈ t • ∧ c ∪ Pt ) ∩ (Te0.2) = φ
(P ∈
∪ T slide 16
∀ t ∈ T , ρ(t) 0
× L) (T × P P × −→ Z∗
m0 : −→ L
otherwise (P × T
otherwise
0
:/DE 92:92+?
92:=92+?
f◦ ◦: :P × T −→ {0, 1}
92+?
)+*3:92+?
f P × T −→ {0, 1}
◦ :*3:92+? × ℘(R) → P × ℘(R) → P
℘(S ) ◦ : ℘(S )
Composition
if A ∈ ℘(S∧ A ∈ •t ∧ → = B ∧ ll ∈ L
if A ∈ ta•set A ∈contexts= and a setL S
◦ : t • ∧ ×/ •t ∧ p PB ∧ S ∈⊆
) of/
℘(R) p
Given
S ⊆ S
2+:92+?
2+:=92+?
otherwise
of context
R ⊆ R ⊆ R

0.2 slide 16
the slide 16


◦ : ℘(S ) × ℘(R) → P

: ℘(S ) ×
0.3 ◦slide 22 ℘(R) → P
contexts
- ) :⊆ S
℘(S) × ℘(R) → P
℘(S S × ℘(R) =(?((#%(!./;,2(.
S

F((G1+
⊆
⊆RS R

0.3 slide 22
92:;$)+

◦({CU , CL }, {∨, CU , CL })

92:=;$)+

Exclusion
-R : ⊆ R (B) and
◦({CU , CL }, {∨, CU , CL })
2+:=;$)+
2+:;$)+
◦({CU , (5)
)+*3:;$)+
*3:;$)+
;$)+
ons(ext(union(S),
R)
union({CU , CL })
relations
0.3 slide 22

union(S)
union(S)
union({CU , CL })
union({CU , CL })
%#!)15
◦({CU , CL }, {∨, CU , CL })
0.3 slide 22
ext∨ ({CU , CL }, {∨, CU , CL })
ext(S, R)
R)
ext(S,

- Add constraints speciﬁc C })
ext∨ ({CR) L }, {∨, CU ,to Leach ◦({CU , CL }, {∨, CU , CL })Pc = Pc ∪ {U|L}
,C
U
cons(S, R)
union({CU , CL })
ext∨ ({CU , CL }, {∨,92:=C2++L })
CU , C
cons(S,dependency relation
92:C2++
context

Pt = Pt ∪ {P r(U|L), P r(¬U|L)}
Pc V Pc ∪ {U|L}
=
Q, CH C

Q, CH ,, CV
Pc = Pc ∪ {U|L}

2+:C2++
2+:=C2++
L })
Pt C, Pt ,∪ V r(U|L), P r(¬U|L)} U , CL }, {∨, CU , Ti = *3:C2++
= CC C {P
C2++
)+*3:C2++

C, CC , CV

Pt = Pt ∪ {P r(U|L), P r(¬U|L)}
Pc = Pc ∪ {U|L}


i = Ti 2013
49

ρ (t) =

C = Pc , ∈ tT Teρ(t),◦ 0◦ ,× T −→, Σ 1} ∗
, , Ti : P ρ, L, m0 {0,
∪ Ti ) = φ
∀ t P : i(P ,×f f, f ∪ (T × P × L) −→ Z
f
T × L)
∗C = P , P , T , T , f, f , ρ, L, m , Σ
) ∪ (T × P × L) −→ Z
m f◦ e P × T ◦−→∗ 0
c 0t : P: ×i L −→ Z {0, 1}
{0, 1}
Pc ∩ Pt = φ
ρ : T −→ Z∗


m0 : P × L −→ Z∗

P∩ P =
ρ : T −→ ∈ ∗Te , ρ(t) = 0
Te c ∩Tit = φ
∀t Z
T ∩ Ti = ∩
φ
∀ t ∈ Te , ∈ T = 0
∀ t ρ(t) i , ρ(t) 0
if A ∈ t •(PcA∪/Pt ) ∧ (T= ∪ Ti )l = L slide 16
∧eA ∈ •t ∧ p = B ∧ 0.2
/ •t p e B ∧ l ∈ φ
∈L
if A ∈ t • ∧ c ∪ Pt ) ∩ (Te0.2) = φ
(P ∈
∪ T slide 16
∀ t ∈ T , ρ(t) 0
× L) (T × P P × −→ Z∗
m0 : −→ L
otherwise (P × T
otherwise
0
:/DE 92:92+?
92:=92+?
f f : ∈ × T −→ {0, 1}
92+?
)+*3:92+?
if A ∈ t • ◦ ◦:AP × T −→ {0, 1} l ∈ L
∧ P •t ∧ p = B ∧
/
◦ :*3:92+? × ℘(R) → P × ℘(R) → P
℘(S ) ◦ : ℘(S )
Composition
if A ∈ ℘(S∧ A ∈ •t ∧ → = B ∧ ll ∈ L
if A ∈ ta•set A ∈contexts= and a setL S
◦ : t • ∧ of/
℘(R) p
otherwise) ×/ •t ∧ p PB⊆∧ S ∈⊆
S
Given
S
2+:92+?
2+:=92+?
otherwise
of context
R ⊆ R ⊆ R

if A ∈ t • ∧ A ∈ •t ∧ p = B ∧ l ∈ L
/

0.2 slide 16
the slide 16
otherwise
◦ : ℘(S ) × ℘(R) → P
: ℘(S ) ×
0.3 ◦slide 22 ℘(R) → P
contexts
- ) :⊆ S
℘(S) × ℘(R) → P
℘(S S × ℘(R) =(?((#%(!./;,2(.0.3 slide 22
92:;$)+
92:=;$)+
F((G1+
⊆
S ⊆RS R
◦({CU , CL }, {∨, CU , CL })
Exclusion
-R ⊆ R (B) and
℘(S ) :× ℘(R) → P
◦({CU , CL }, {∨, CU , CL })
2+:=;$)+
2+:;$)+
◦({CU , (5)
)+*3:;$)+
*3:;$)+
;$)+
ons(ext(union(S),
R)
union({CU , CL })
relations

0.3 slide 22

union(S)
union(S)
union({CU , R),
cons(ext(union(S),CL })R)
0.3 slide 22
ext(S, R)
ext(S, R)
union(S)

union({CU , CL })
%#!)15

◦({CU , CL }, {∨, CU , CL })
ext∨ ({CU , CL }, {∨, CU , CL })

- Add constraints speciﬁc C })
ext∨ ({CR) L }, {∨, CU ,to Leach ◦({CU , CL }, {∨, CU , CL })Pc = Pc ∪ {U|L}
,C
cons(S, R)
union({CU , CL })
ext∨ ({CU , CL }, {∨,92:=C2++L })
CU , C
cons(S,dependency relation
ext(S, U
92:C2++
context

Pt = Pt ∪ {P r(U|L), P r(¬U|L)}
Pc V Pc ∪ {U|L}
=
Q, CH C

cons(S, R)
Q, CH ,, CV
Pc = Pc ∪ {U|L}

2+:C2++
2+:=C2++
L })
Pt-Q,Pt ,∪ V r(U|L), P r(¬U|L)} U , CL }, {∨, CU , Ti = *3:C2++
= C constraints applicable for all
AddC C
C2++
C, C CH ,C {P
)+*3:C2++
CV

C, C , V

Pt = Pt ∪ {P r(U|L), P r(¬U|L)}
Pc = Pc ∪ {U|L}
deﬁned contexts
TMonday 8 July∪ C

i = Ti C2013V
49
C, C , {act(U|L), deac(U|Ln )}

ρ (t) =

Consistent States
Consistent States
No enabled internal transitions, and no marked
temporary places
D#%%67!+$0/(8-9(E!;$)+!#(%*!92+?F

*3:;$)+

92:=!;$)+

;$)+
2+:=;$)+

2+:;$)+ 92:;$)+

92:92+?

2+:92+?

Monday 8 July 2013

)+*3:;$)+

)+*3:92+?

*3:92+?

92+?

2+:=92+?

92:=92+?

50

Consistent States
Consistent States
No enabled internal transitions, and no marked
temporary places
D#%%67!+$0/(8-9(E!;$)+!#(%*!92+?F

✓

*3:;$)+

✓

92:=!;$)+

;$)+
2+:=;$)+

2+:;$)+ 92:;$)+

92:92+?

2+:92+?

Monday 8 July 2013

)+*3:;$)+

)+*3:92+?

*3:92+?

92+?

2+:=92+?

92:=92+?

50

Activation Semantics of CoPNs
92:=;$)+

*3:;$)+
2+:;$)+ 92:;$)+

;$)+

2+:=;$)+

)+*3:;$)+

%#!)F/%5
92:9$%

2+:9$%

)+*3:9$%

*3:9$%

9$%

2+:=9$%

92:=9$%

Context Petri Nets: Consistent of Context-dependent Behavior.
Cardozo, Nicolás and Vallejos, Jorge and González, Sebastián and Mens, Kim
and D’Hondt,Theo. International Workshop on Petri Nets and Software
Engineering (PNSE ‘12).
Monday 8 July 2013

51

92:=;$)+

*3:;$)+
2+:;$)+ 92:;$)+

;$)+

2+:=;$)+

)+*3:;$)+

81%)'1)(:;,2(

%#!)F/%5
92:9$%

2+:9$%

)+*3:9$%

*3:9$%

9$%

2+:=9$%

92:=9$%

1. External transitions ﬁre after G*3-8*+ and G)+*3-8*+

Monday 8 July 2013

51

92:=;$)+

*3:;$)+
2+:;$)+ 92:;$)+

;$)+

2+:=;$)+

)+*3:;$)+

81%)'1)(:;,2(

%#!)F/%5
92:9$%

2+:9$%

)+*3:9$%

*3:9$%

9$%

2+:=9$%

92:=9$%

2. Enabled internal transitions are always ﬁred

Monday 8 July 2013

51

92:=;$)+

*3:;$)+
2+:;$)+ 92:;$)+

;$)+

2+:=;$)+

)+*3:;$)+

81%)'1)(:;,2(

%#!)F/%5
92:9$%

2+:9$%

)+*3:9$%

*3:9$%

9$%

2+:=9$%

92:=9$%

3. If there is an inconsistency, actions revert to previous consistent state
Monday 8 July 2013

51

92:=;$)+

*3:;$)+
2+:;$)+ 92:;$)+

;$)+

2+:=;$)+

)+*3:;$)+

81%)'1)(:;,2(
81%)'1)(:=,!

%#!)F/%5
92:9$%

2+:9$%

)+*3:9$%

*3:9$%

9$%

2+:=9$%

92:=9$%

3. If there is an inconsistency, actions revert to previous consistent state
4. Consistent states are accepted
Monday 8 July 2013

51

Context Activation Simulation Tool
H,#)(9)!I
=(?((#%(!
;,2(
F((G1+
H,#)(9)/2(J(#2(#%K/(1)',#!I
;,2(/LM/=(?((#%(!
F((G1+/NL/;,2(

Monday 8 July 2013

52

Context Activation Simulation Tool

Monday 8 July 2013

52

Petri net analyses
“Unfold” the CoPN to a traditional Petri net
need to introduce bounds = some loss of semantics

92:I

*3:I

I

2+:=I

92:=I

)+*3:I

2+:I

%#!)85

%#!)G5

2+:H 92:H

Monday 8 July 2013

*3:H

H

2+:=H

92:=H

)+*3:H

53

0.8 Slide 18

Petri net analyses


1







1

f (t, p, l) =




1




f (p, t, l)

if p = Q ∧ M ∈ •t ∧ M ∈ •t ∧
/
Q ∈ ◦t ∧ l ∈ L
/
if p = P r(¬Q) ∧ M ∈ •t ∧ M ∈ t • ∧
/
Q ∈ ◦t ∧ l ∈ L
/
if p = P r(Q) ∧ M ∈ t • ∧ M ∈ •t ∧ l ∈ L
/
otherwise

need to introduce bounds = some loss of semantics
92:I

*3:I

I

2+:=I

92:=I

)+*3:I

0.8 Slide 18

2+:I

%#!)85

%#!)G5

2+:H 92:H

*3:H

H

2+:=H

1

92:=H

)+*3:H

req(A)

Pr(A)

A1

act(A)0

req(¬A)

Pr(¬A)

deac(A)1

A0
deac(A)2

1

deac(B)1 deac(B)2

deac(A)1
1

B0
act(B)0
req(B)

Pr(B)

Monday 8 July 2013

1

B1
act(B)1

deac(B)1

1

B2

req(¬B)

Pr(¬B)

deac(B)2

53

Petri net analyses
Use a traditional Petri net analysis tool like LoLA
Choose appropriate bounds to avoid too large Petri nets
Use LoLa to reason about properties like reachability and liveness
partly manual (guided by the user)
partly automated (analysis tests generated from
dependencies)
Monday 8 July 2013

54

Conclusion
2

1
f (p, t, l) =
f (p, t, l)

1
f (t, p, l) =
f (t, p, l)
requirement

implication
0.4 slide 16
exclusion

disjunction

Formalization Formal basis for adaptations
if A ∈ t • ∧ A ∈ •t ∧ p = B ∧ l ∈ L
/
and their interactions
otherwise

if A ∈ t • ∧ A ∈ •t ∧ p = B ∧ l ∈ L
/
otherwise

CoPN model
conjunction

suggestion
causality

◦ : ℘(S ) × ℘(R) → P

S ⊆ S
R ⊆ R

◦(S, R) = cons(ext(union(S), R), R)

CoPNs composition
union(S)
ext(S, R)

Analysis

cons(S, R)
Q, CH , CV
C, CC , CV

0.5 slide 17

Execution

Monday 8 July 2013

External transition firing: occurs only at the beginning of a step Υ when
P is in a consistent state m (i.e., the marking of P is consistent) and E is empty
˜
(i.e., i.e. no internal transitions are enabled). After ﬁring an external transition,
the marking of the COPN! (COPN!) is modiﬁed, possibly enabling transitions
in Ti . Such transitions become the elements in the priority set E.
t ∈ Te , m[tm
˜

55

Conclusion
2

1
f (p, t, l) =
f (p, t, l)

1
f (t, p, l) =
f (t, p, l)
0.1 slide 14

otherwise

if A ∈ t • ∧ A ∈ •t ∧ p = B ∧ l ∈ L
/
otherwise

CoPN model

1

requirement
slide 14

disjunction

implication
0.4 slide 16
0.1

if A ∈ t • ∧ A ∈ •t ∧ p = B ∧ l ∈ L
/

C = Pc , Pt , Te , Ti , f, f◦ , ρ, L, m0 , Σ

Pc ∩ Pt = φ
Te ∩ Ti = φ
(Pc ∪ Pt ) ∩ (Te ∪ Ti ) = φ
f : (P × T × L) ∪ (T × P × L) −→ Z∗
f◦ : P × T −→ {0, 1}

suggestion

ρ : T −→ Z∗
∀ t ∈ Te , ρ(t) = 0
∀ t ∈ Ti , ρ(t) 0
m0 : P × L −→ Z∗

exclusion

conjunction

causality

◦ : ℘(S ) × ℘(R) → P

0.2 slide 15

S ⊆ S
slide R
R 0.3⊆ 16
∧

S ⊆ S
R ⊆ R

◦ : ℘(S ) × ℘(R) → P

union(S)

Execution model
for COP systems

CoPNs composition
union(S)

ext(S, R)

2

2



cons(S, R)

0.4 slide 17

t ∈ E, ∈ T , m[tm
/
External transition firing: toccurs only at the beginning of a step Υ when
(0.2)
PP, E, T , T , m → [m/m ]P, enabledmof P ), T t, t, E, and E˜is empty
is in a consistent state m (i.e., the marking (Ti is consistent) m, m
˜
˜
Evaluation of the COPN! (COPN!) is modified, possibly enabling transitions
the marking termination: When there are no more internal transitions to
t cases ∈ possible:
/
0.6be fired22 the set E, three∈ E, the elements in the priority set E.
slide in
in Ti . Such transitions become tare T , m[tm
(0.2) 3
1. The , T m occurs when for m marking multiset of P there are marked
P, E, T first,case → [m/m ]P, enabledm (Ti ), T t, t, E, m, m
˜
˜
t ∈ Te m[tm
˜
temporary places and not all initially, enabled internal transitions have(0.1)
been
2
]P, enabled more internal transitions to
˜
˜
Evaluation termination: → [ ˜ there are no mthe), φ, φ, m of COPN! m
fired. such a P, φ, all changes are rolled back (Ti state of
case all m When
fired. InIn such acase φ, φ,changes m/m rolledback toto the state the the COPN! m
are
besaved before transition firing: If possible: is not empty, firing one of the
fired in the the firing of the first internal transition
Internal set E, three cases are the set E
saved before the firing of the first internal transition t: t:
1.internal transitions with when for m markinga multiset of Pm of the COPN!
The first case occurs highest priority yields new marking there are marked
markedm (P initially T , m[tm π (T transitions have to
temporary placesofand not (Pttbeforet firing ππ(T ), ), 2π2is) saved T a means been
P. The state markedm allt ) ∈ E, m enabled internal T (as
the COPN! ) = φ, ∈ = transition t (T )
= φ, / = 3 3 (T
m
(0.3)
(0.2)
, m T T ˜
˜
state
t,
˜
facilitate P, E, T , T all φ,→ ,[m/m inconsistencies). ),Tthe newE, of m possibly (0.3)
backtracking changesof →enabledm (TtoThe m marking
fired. In such a caseP,φ,in case, m]P,rolled/m]P, φ,φ, ,T ,t,˜ mm,the COPN! m
are [m back i T φ, φ, ˜
P, T , T , m → [m /m]P,
˜
enables some internal occurs when Tforwhich become the
saved before second case transitions internalm marking t: elements P there are
the firing of the first in i , transition multiset of of E.
2. The
Evaluation termination: When there are no more internal transitions to

2. The second the set occurs cases arefor m initially enabled haveof P there are
temporaryfired in case E, three all transitions marking multiset been fired,
be places marked and when possible:
(T
the COPN! is rolledcase occurs = φ,for m marking multiset of T
temporary 1. The markedm (Ptitswhen transitions state 2˜ andPthe firing marked (0.3)
placesfirst back to ) all consistent initially ) there are request is
marked and last m = π3 (T ), π m enabled have been fired,

temporary places and
P,
signaled as rolled φ, , not all →consistent internal m the firing request is
the COPN! isdenied: backTtoT , m initially enabled state, mtransitions have been
its ˜
last [m /m]P, φ, T φ, and
˜ ˜
fired. In such a case all changes are rolled back to the state of the COPN! m
signaled as second case firing of the first )internalπ markingt: multiset of P there are
2. The denied: the occurs when for φ, transition T
saved before
marked (P = m (T )

ext(S, R)

Analysis

cons(S, R)
Q, CH , CV
C, CC , CV

2

temporary places marked T , T , mtransitionsm]P, φ, φ, enabled have been(0.4)
fired,
P,markedmallt)(P φ, = φ,/ 3 initially φ, T
φ, and (P = [m π ˜(T(T π2 (T T m
˜
markedm → m = π2 ), ⊆ ) ˜
t
(0.3)
the COPN! is rolled backP, φ, T ,lastm)→ [m /m]P, φ,)Tm and the firing request (0.4)
to its T consistent state ˜φ, m
is
, ˜
3. The third case φ, T , when , ˜→ marking multiset m
˜
˜
˜
signaled as denied: P,occurs T , mfor m[m /m]P, φ, φ, φ, of P no temporary

m

t

0.5 slide 17

2. The second case occurs when for m marking multiset of P there m
places remain marked, we finish the step by turning the current markingare of

marked and
3. Thetemporary the occurs when all)transitions(T ) multisethaveP nofired,
third case new consistentt state m:
m marking enabled of been temporary
the COPN! into places markedm (Pfor= φ,˜ π2 initiallyT
the COPN! is rolled
˜
places remain marked, weback ,to its last consistent state m andm firing request is (0.4) of
finish the step by˜turning the the
current marking m
˜
signaled as P, φ, T , T m → [m ) m]P, φ, φ, φ, ˜
denied:
markedm (P / =
the COPN! into the new consistent state tm: φ
˜
(0.5)

P, φ, markedm m = φ, πφ, φ, φ, m
T , T m ) P, 2 (T T
˜
3. The third case occurs when ,for(Pt→ marking) multiset of P no temporary
(0.4)
places remain marked, we finish T , m → (Ptturning φ, φ, m
current marking m of
P, φ, T markedm by ) m]P, φ, the ˜
, the step [m / = φ
˜
˜
the COPN!3. Thethe new consistent m →˜ markingφ, φ, m P no temporary (0.5)
into third case occurs when
state
0.5 slide 22 P, φ, T , T , ˜ for mm:P, φ, multiset of

Execution

CoPN semantics
0.6 slide 22
P is in a22consistent state m (i.e., the marking of P is consistent) and E is empty
˜
0.5 slide
the
CoPNs marking of the COPN! (COPN!) is modified, possibly enabling transitions
places remain marked, we finish the step by turning the current marking m of
markedm (Pt m:
the COPN! into the new consistent {∨, C˜=Cφ})
◦({CU , CL }, state ) U , L

P, φ, T , T , marked (P ) = φ φ, m
m → P, φ, φ,
˜
m

0.5 slide 22

t

P, φ, T , T , m → CP, φ, φ, φ, m
˜
union({CU , L })

(0.5)

(0.5)

◦({CU , CL }, {∨, CU , CL })

ext∨ ({CU , },L{∨, C CU , CL })
◦({C , C C }, {∨, , C })
U

L

U

L

union({CU , CL CL
◦({CU , CL }, {∨, CU ,}) })

Pc = Pc ∪ {U|L}

Pt = Punion({CU , C,LCLr(¬U|L)}
union({CU })
t ∪ {P r(U|L), P })

ext∨ ({CU , CL }, {∨, CU , CL })

Ti = Ti ∪ {act(U|L), deac(U|Ln )}
ext∨({CU ,P L }, {∨, CU , C })

Pc C C ∪ {U|L}
ext∨ ({CU , =L },c{∨, CU , CLL})

ρ (t) = P = P Pc = Pc ∪ {U|L}
∪ = r(U|L), ∨
Pif t {Pact(U|L)P r(¬U|L)}
t c = Pc ∪ {U|L} t = deac(U|L)
t2
P = P ∪ {P r(U|L), P r(¬U|L)}
t

t

i
Tt T= T {act(U|L),P r(¬U|L)}
Pi =TPt ∪ i{P{act(U|L), deac(U|Ln )}
∪ r(U|L), deac(U|Ln )}
i

Monday 8 July 2013

Ti = Tiρ(t){act(U|L), deac(U|Ln )}
∪ if t inTe ∪ Ti
ρ (t)ρ(t) if t inTe ∪ Ti
ρ (t) = = 2
if t = act(U|L) ∨ t = deac(U|L)
2
= e ∪ Ti
ρ(t) ifift tinTact(U|L) ∨ t = deac(U|L)
ρ =
(t)
2

t ∈ Te , m[tm
˜

55

Conclusion
2

1
f (p, t, l) =
f (p, t, l)

1
f (t, p, l) =
f (t, p, l)
0.1 slide 14

otherwise

if A ∈ t • ∧ A ∈ •t ∧ p = B ∧ l ∈ L
/
otherwise

CoPN model

1

requirement
slide 14

disjunction

implication
0.4 slide 16
0.1

if A ∈ t • ∧ A ∈ •t ∧ p = B ∧ l ∈ L
/


Pc ∩ Pt = φ
Te ∩ Ti = φ
(Pc ∪ Pt ) ∩ (Te ∪ Ti ) = φ
f : (P × T × L) ∪ (T × P × L) −→ Z∗
f◦ : P × T −→ {0, 1}

0.8 Slide 18

causality

◦ : ℘(S ) × ℘(R) → P

f (t, p, l) =

0.2 slide 15

S ⊆ S
slide R
R 0.3⊆ 16

0.8 Slide 18
◦ : ℘(S ) × ℘(R) → P

CoPNs composition
union(S)

ext(S, R)

2

Execution model
for COP systems


union(S)

2


1







1

Reasoning and analysis
over dynamic adaptations





1




f (p, t, l)

if p = Q ∧ M ∈ •t ∧ M ∈ •t ∧
/
Q ∈ ◦t ∧ l ∈ L
/
if p = P r(¬Q) ∧ M ∈ •t ∧ M ∈ t • ∧
/
Q ∈ ◦t ∧ l ∈ L
/
if p = P r(Q) ∧ M ∈ t • ∧ M ∈ •t ∧ l ∈ L
/
otherwise

Reasoning module

∧

S ⊆ S
R ⊆ R

5

suggestion

ρ : T −→ Z∗
∀ t ∈ Te , ρ(t) = 0
∀ t ∈ Ti , ρ(t) 0
m0 : P × L −→ Z∗

exclusion

conjunction

cons(S, R)

0.4 slide 17

t ∈ E, ∈ T , m[tm
/
(0.2)
˜
˜
/
slide in
(0.2) 3
˜
˜
t ∈ Te m[tm
˜
been
2
˜
˜
case all m When
are
= φ, / = 3 3 (T
m
(0.3)
(0.2)
, m T T ˜
˜
state
t,
˜
P, T , T , m → [m /m]P,
˜
2. The

(T

P,
its ˜
˜ ˜
saved before
marked (P = m (T )

1

req(A)

Q, CH , CV

A1

act(A)0

req(¬A)

Pr(¬A)

deac(A)1

A0
deac(A)2

1

ext(S, R)

cons(S, R)

Pr(A)

Analysis

deac(B)1 deac(B)2

deac(A)1
1

B0
act(B)0
req(B)

Pr(B)

B1
act(B)1

deac(B)1

1

1

req(¬B)

Pr(¬B)

deac(B)2

B2

C, CC , CV

2

ﬁred,
˜
markedm → m = π2 ), ⊆ ) ˜
t
(0.3)
is
, ˜
˜
˜
˜

m

t

0.5 slide 17


marked and
the COPN! is rolled
˜
current marking m
˜
denied:
markedm (P / =
˜
(0.5)

T , T m ) P, 2 (T T
˜
(0.4)
˜
˜
state

Execution

CoPN semantics
0.6 slide 22
˜
0.5 slide
the
markedm (Pt m:

m → P, φ, φ,
˜
m

0.5 slide 22

t

P, φ, T , T , m → CP, φ, φ, φ, m
˜
union({CU , L })

(0.5)

(0.5)

◦({CU , CL }, {∨, CU , CL })

ext∨ ({CU , },L{∨, C CU , CL })
◦({C , C C }, {∨, , C })
U

L

U

L

union({CU , CL CL
◦({CU , CL }, {∨, CU ,}) })

Pc = Pc ∪ {U|L}

union({CU })
t ∪ {P r(U|L), P })

ext∨ ({CU , CL }, {∨, CU , CL })

ext∨({CU ,P L }, {∨, CU , C })

Pc C C ∪ {U|L}
ext∨ ({CU , =L },c{∨, CU , CLL})

ρ (t) = P = P Pc = Pc ∪ {U|L}
∪ = r(U|L), ∨
t2
P = P ∪ {P r(U|L), P r(¬U|L)}
t

t

i
i

Monday 8 July 2013

ρ (t) = = 2
2
= e ∪ Ti
ρ =
(t)
2

t ∈ Te , m[tm
˜

55

Conclusion
2

1
f (p, t, l) =
f (p, t, l)

1
f (t, p, l) =
f (t, p, l)
0.1 slide 14

otherwise

if A ∈ t • ∧ A ∈ •t ∧ p = B ∧ l ∈ L
/
otherwise

CoPN model

1

requirement
slide 14

disjunction

implication
0.4 slide 16
0.1

if A ∈ t • ∧ A ∈ •t ∧ p = B ∧ l ∈ L
/


Pc ∩ Pt = φ
Te ∩ Ti = φ
(Pc ∪ Pt ) ∩ (Te ∪ Ti ) = φ
f : (P × T × L) ∪ (T × P × L) −→ Z∗
f◦ : P × T −→ {0, 1}

0.8 Slide 18

causality

◦ : ℘(S ) × ℘(R) → P

f (t, p, l) =

0.2 slide 15

S ⊆ S
slide R
R 0.3⊆ 16

0.8 Slide 18
◦ : ℘(S ) × ℘(R) → P

CoPNs composition
union(S)

ext(S, R)

2

Execution model
for COP systems


union(S)

2


1







1

Reasoning and analysis
over dynamic adaptations





1




f (p, t, l)

if p = Q ∧ M ∈ •t ∧ M ∈ •t ∧
/
Q ∈ ◦t ∧ l ∈ L
/
if p = P r(¬Q) ∧ M ∈ •t ∧ M ∈ t • ∧
/
Q ∈ ◦t ∧ l ∈ L
/
if p = P r(Q) ∧ M ∈ t • ∧ M ∈ •t ∧ l ∈ L
/
otherwise

Reasoning module

∧

S ⊆ S
R ⊆ R

5

suggestion

ρ : T −→ Z∗
∀ t ∈ Te , ρ(t) = 0
∀ t ∈ Ti , ρ(t) 0
m0 : P × L −→ Z∗

exclusion

conjunction

cons(S, R)

0.4 slide 17

t ∈ E, ∈ T , m[tm
/
(0.2)
˜
˜
/
slide in
(0.2) 3
˜
˜
t ∈ Te m[tm
˜
been
2
˜
˜
case all m When
are
= φ, / = 3 3 (T
m
(0.3)
(0.2)
, m T T ˜
˜
state
t,
˜
P, T , T , m → [m /m]P,
˜
2. The

(T

P,
its ˜
˜ ˜
saved before
marked (P = m (T )

1

req(A)

Q, CH , CV

A1

act(A)0

req(¬A)

Pr(¬A)

deac(A)1

A0
deac(A)2

1

ext(S, R)

cons(S, R)

Pr(A)

Analysis

deac(B)1 deac(B)2

deac(A)1
1

B0
act(B)0
req(B)

Pr(B)

B1
act(B)1

deac(B)1

1

1

req(¬B)

Pr(¬B)

deac(B)2

B2

C, CC , CV

2

ﬁred,
˜
markedm → m = π2 ), ⊆ ) ˜
t
(0.3)
is
, ˜
˜
˜
˜

m

t

0.5 slide 17


marked and
the COPN! is rolled
˜
current marking m
˜
denied:
markedm (P / =
˜
(0.5)

T , T m ) P, 2 (T T
˜
(0.4)
˜
˜
state

Execution

CoPN semantics
0.6 slide 22
˜
0.5 slide
Simulation of interactions
the
Simulation tool
in Ti . Such transitions become the elements in the priority set E. between contexts
markedm (Pt m:

m → P, φ, φ,
˜
m

0.5 slide 22

t

P, φ, T , T , m → CP, φ, φ, φ, m
˜
union({CU , L })

(0.5)

(0.5)

◦({CU , CL }, {∨, CU , CL })

ext∨ ({CU , },L{∨, C CU , CL })
◦({C , C C }, {∨, , C })
U

L

U

L

union({CU , CL CL
◦({CU , CL }, {∨, CU ,}) })

Pc = Pc ∪ {U|L}

union({CU })
t ∪ {P r(U|L), P })

ext∨ ({CU , CL }, {∨, CU , CL })

ext∨({CU ,P L }, {∨, CU , C })

Pc C C ∪ {U|L}
ext∨ ({CU , =L },c{∨, CU , CLL})

ρ (t) = P = P Pc = Pc ∪ {U|L}
∪ = r(U|L), ∨
t2
P = P ∪ {P r(U|L), P r(¬U|L)}
t

t

i
i

Monday 8 July 2013

ρ (t) = = 2
2
= e ∪ Ti
ρ =
(t)
2

t ∈ Te , m[tm
˜

55

Research @ RELEASeD (presented at SATTOSE2013)

Recommended

Recommended

More Related Content

Viewers also liked

Viewers also liked (8)

Similar to Research @ RELEASeD (presented at SATTOSE2013)

Similar to Research @ RELEASeD (presented at SATTOSE2013) (20)

More from kim.mens

More from kim.mens (20)

Recently uploaded

Recently uploaded (20)

Research @ RELEASeD (presented at SATTOSE2013)