Explanations and error diagnosis

Description

Explanations and error diagnosis LIFO Gerard Ferrand, Willy Lesaint, Alexandre Tessier public, rapport de recherche D3.2.2 Contents 1 Introduction 3 2 Preliminary notations and denitions 4 2.1 Notations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 2.2 Constraint Satisfaction Problem . . . . . . . . . . . . . . . . . . . . . 4 2.3 Constraint Satisfaction Program . . . . . . . . . . . . . . . . . . . . . 5 2.4 Links between CSP and program . . . . . . . . . . . . . . . . . . . . 7 3 Expected Semantics 8 3.1 Correctness of a CSP . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 3.2 Symptom and Error . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 4 Explanations 9 4.1 Explanations . . . . . . . . . . . . . . . . . . . . . . . . .

• local consis


• traction algorithms


• symptom can


• dynamic constraint


• nite domains using


• constraint programming


• declarative diagnosis


• domain reduction only


• constraint satisfaction

Sujets

• Ferrand
• Tessier
• Constraint programming
• Constraint satisfaction

Informations

• Publié par: pefav
• Nombre de lectures: 10
• Langue: English

Extrait

.
Explanations
.
and
.
error
.
diagnosis
.
LIF
.
O
.
G
.

.
erard
Symptom
F
.
errand,
.
Willy
and
Lesain
.
t,
9
Alexandre
.
T
.
essier
.
public,
12
rapp
.
ort
.
de
.
rec
.
herc
.
he
.
D3.2.2
.
Con
.
ten
.
ts
.
1
.
In
.
tro
.
duction
.
3
Computed
2
.
Preliminary
.
notations
.
and
12
denitions
.
4
.
2.1
.
Notations
13
.
.
.
.
.
.
.
6
.
.
.
.
.
.
.
3.2
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4
.
Explanations
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
10
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Error
4
F
2.2
Error
Constrain
.
t
.
Satisfaction
.
Problem
.
.
.
.
.
.
Diagnosis
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
14
.
.
.
.
.
.
.
.
.
.
.
.
.
8
.
Symptom
.
Error
.
.
4
.
2.3
.
Constrain
.
t
.
Satisfaction
.
Program
.
.
.
.
.
.
.
.
.
.
.
.
.
.
8
.
Explanations
.
4.1
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
5
.
2.4
.
Links
.
b
.
et
4.2
w
explanations
een
.
CSP
.
and
.
program
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
5
.
Diagnosis
.
5.1
.
rom
.
to
.
.
.
.
.
.
.
.
.
.
.
.
7
.
3
.
Exp
.
ected
.
Seman
.
tics
.
8
.
3.1
5.2
Correctness
Algorithms
of
.
a
.
CSP
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
13
.
Conclusion
.
1
.an
Abstract
ma
The
2
rep
e
ort
debugging,
prop
of
oses
of
a
diagnosis
theoretical
k
approac
erator,
h
alues
of
remo
the
considered
debugging
They
of
com-
constrain
cating
t
b
programs
declarativ
based
cal
on
closure,
the
of
notion
con-
of
v
explanation
als.
tree
b
(D1.1.1
the
and
t
D1.1.2
a
part
of
2).
trace.
The
in
prop
error
osed
ro
approac
a
h
ords
is
diagnosis,
an
,
attempt
tency
to
oin
adapt
e
algorithmic
als
debugging
v
to
as
constrain
sequence
t
other
programming.
alue
In
v
this
Explanations
theoretical
y
framew
e
ork
as
for
essence
domain
constrain
reduction,
programming.
explanations
are
are
declarativ
pro
view
of
the
trees
putation
explaining
The
v
consists
alue
lo
remo
an
v
in
als.
explanation
These
oted
pro
y
of
symptom.
trees
eyw
are
:
dened
e
b
algorithmic
y
CSP
inductiv
lo
e
consis-
denitions
op
whic
x-p
h
t,
express
inductiv
the
denition
remo
vcall
1
e
In
minimal
tro
duction
duction
in
Declarativ
a
e
b
diagnosis
a
[15
here
]
Strictly
(also
symptom
kno
family
wn
o
as
eling
algorithmic
as
debugging)
debugging
ha
anomaly
v
answ
e
the
b
en,
een
input
success-
wledge
fully
pro
used
].
in
man
dieren
.
t
[12
programming
prop
paradigms
necessary
(e.g.
in
logic
their
programming
in
[15
dur-
],
cte
func-
.
tional
wrong
programming
for
[10]).
y
De
erroneous
clar
ts
ative
a
means
actual
that
exp
the
the
user
already
has
correctness
no
re-
need
of
to
v
consider
constrain
the
dynamic
computational
er,
b
failure
eha
tro
vior
has
of
But
the
and
programming
can
system,
The
he
pro
only
of
needs
used
a
programming.
declarativ
oin
e
earance
kno
of
wledge
to
of
erties
the
e
exp
b
ected
missing
prop
eals
erties
(a
of
er
the
Symptoms
program.
c
This
lo
pap
a
er
diagnosis
is
for
an
explanation
attempt
diagnostic
to
at
adapt
am
declarativ
kno
e
seman
diagnosis
giv
to
alue
constrain
It
t
ermitted
programming
e
thanks
a
to
constrain
a
algorithms
notion
closed
of
ha
explanation
een
tree.
useful
Constrain
applications:
t
satisfaction
programs
er-constrained
are
ktrac
not
Moreo
easy
has
to
used
debug
ysis
b
The
ecause
of
they
the
are
b
not
in
algorithmic
in
programs
the
[14]
not
and
(lab
tracing
e
tec
constrain
hniques
de-
are
pap
rev
us
ealed
e
limited
computation
in
structure
fron
adapt
t
constrain
of
rom
them.
e
More-
w
o
the
v
an
er
the
it
program.
w
relativ
ould
exp
b
pr
e
the
incoheren
an
t
d
to
symptom
use
a
only
or
lo
er.
w
er
lev
lac
el
constrain
debugging
constrain
to
This
ols
cuses
whereas
answ
for
caused
these
one
languages
aints
the
eaking,
emphasis
of
is
t,
on
is
declarativ
err
e
It
seman
searc
tics.
kind
Here
in
w
F
e
clar
are
,
in
ust
terested
(1)
in
o
a
(2)
wide
(3)
eld
of
of
cte
applications
This
of
b
constrain
b
t
v
programming:
from
nite
domain".
domains
has
and
p
pr
to
op
v
agation
the
.
of
The
large
aim
of
of
t
constrain
traction
t
[6
programming
A
is
notion
to
explanations
solv
v
e
b
Constrain
pro
t
ed
Satisfaction
in
Problems
y
(CSP)
dynamic
[17],
t
that
problems,
is
v
to
problems,
pro
bac
vide
king,.
an
.
instan
v
tiation
it
of
also
the
een
v
for
ariables
anal-
whic
in
h
].
is
in
solution
duction
of
lab
the
in
constrain
formalism
ts.
already
The
een
solv
osed
er
[13].
go
this
es
tro
to
complicates
w
formalism
ards
is
the
really
solutions
here
com
eling
bining
b
t
considered
w
additional
o
ts).
dieren
explanations
t