La lecture en ligne est gratuite
Le téléchargement nécessite un accès à la bibliothèque YouScribe
Tout savoir sur nos offres
Télécharger Lire

Proof planning with multiple strategies [Elektronische Ressource] / Andreas Meier

283 pages
ersitProhenofkPlanningatwithSaarlandesMultipleecStrategiesakultAndreasderMeieratDissertationzur2004ErlangungdesFGradesDoktorIderUnivhaftendes(Dr.-Ing.)Saarbrderucen,h-Ta,DekenanProf.Prof.2.Dr.PhilippSlusallek,desUnivProf.ersitat3.atorgdesSaarbrSaarlandes,UnivSaar-ucbrhGertucersitkSaarlandes,enkVhorsitzenderJProf.UnivDr.desRaimucundquiumSeidel,UnivSaarbrersitenatterdesDr.Saarlandes,SmolkSaar-UnivbrdesucSaarbrkucenen1.terhDr.terPD.Siekmann,Dr.ersitEricaatMelis,Saarlandes,UniversitkKolloat06.02.2004,desersitSaarlandes,atSaarbrSaarlandes,uckkentiThisknothesistationpresendomaintsplanningproplanningofwplanningThewithpromimplemenultiplestrategiesstrategies.aluationStrategiesstrategiesareisindepten-discusseddenthettheproofWplanproopmerations,theandThedierenprotmstrategiesitsrealizeMuldierenwithtlargeplanorenementhattsthisasstudieswortanceellwledgeaselplanplanning.modomaindications.wledge.Comparedewithtedtheofpreviouswithproultipleofinplanning,Mulmsystem.ulti-evpleofstrategyofprowithofultipleplanningandinimplementrointianothertolevandelwandsmalleritsstudiesareintrol.thesis.
Voir plus Voir moins

ersit
Pro
hen
of
k
Planning
at
with
Saarlandes
Multiple
ec
Strategies
akult
Andreas
der
Meier
at
Dissertation

zur
2004
Erlangung

des
F
Grades

Doktor
I
der
Univ


haften
des
(Dr.-Ing.)
Saarbr
der
uc

en,

h-Ta,
Dek
en
an
Prof.
Prof.
2.
Dr.

Philipp

Slusallek,
des
Univ
Prof.
ersit
at

3.
at
org
des
Saarbr
Saarlandes,
Univ
Saar-
uc
br
h

Gert
uc
ersit
k
Saarlandes,
en
k
V
h
orsitzender
J
Prof.
Univ
Dr.
des
Raim
uc
und
quium
Seidel,

Univ
Saarbr
ersit
en


at
ter
des
Dr.
Saarlandes,
Smolk
Saar-
Univ
br


des
uc
Saarbr
k
uc
en
en
1.


ter
h
Dr.
ter

PD.
Siekmann,
Dr.
ersit
Erica
at
Melis,
Saarlandes,
Univ

ersit
k

Kollo
at
06.02.2004,
des
ersit
Saarlandes,
at
Saarbr
Saarlandes,


uc
k
k
en

ti
This
kno
thesis
tation
presen
domain
ts
planning
pro
planning
of
w
planning
The
with
pro
m
implemen
ultiple
strategies
strategies.
aluation
Strategies
strategies
are
is
indep
t
en-
discussed
den
the
t
the
pro

of
W
plan
pro
op
m
erations,
the
and
The
dieren
pro
t
m
strategies
its
realize
Mul
dieren
with
t
large
plan
o
renemen
that
ts
this
as
studies
w
ortance
ell
wledge
as
el
plan
planning.
mo
domain
dications.
wledge.
Compared
e
with
ted
the
of
previous
with
pro
ultiple
of
in
planning,
Mul
m
system.
ulti-
ev
ple
of
strategy
of
pro
with
of
ultiple
planning
and
in
implemen
tro
in

ti
another


t

o
lev
and
el
w
and
smaller
its
studies

are

in
trol.
thesis.
Both,

the
illustrate
strategies
imp
and
of
the
kno

at

strategy-lev
trol
for

of

demit
Kurzzusammenfassung
(mathematisc
Diese
eransc
Arb
mehreren
eit
durc
stellt
v
Bew

eisplanen
System.
mit
ei
mehreren
eit
Strategien
das
v
b
or.
ub
Strategien
dieren.
sind
tiert
un-
v
abh
mit

ei
angige
die
Komp
Die
onen
Dom
ten
ene
f
es

erden
ur
Wissen
das
eine
Bew
k
eisplanen,
eisplanen
w
ist
ob
Mul
ei
Ev
v
Bew
ersc
Strategien
hiedene
ti
Strategien
und
v
F
er-

sc
dieser
hiedene
w
V
allstudien
erfeinerungen
hen
o
anen
der
der
Mo
on
dik
und
ationen
Bew
eines
utzt
Bew
ann.
eisplans
hes)
realisieren

k
er

Dom
onnen.
ane
Im
o
V
Bew

mit
h
Strategien
mit
implemen
dem
im
bisherigen
ti
Bew
Zur
eisplanen
aluierung
f
on

eisplanen
uhrt
mehreren
Bew
wurden
eisplanen
Mul
mit
zw
mehreren
groe
Strategien
zw
eine
kleinere
neue
allstudien

hgef
hieeb
uhrt,
ene
in
und
Arb
deren
diskutiert
heuristisc
erden.
he
F
Kon
v
trolle
haulic
ein.
das
So

w
wissen,
ohl
auf
die
Eb
Strategien
v
selbst
Strategien
als
orliegt,

wie
h
im
ihre
eisplanen
Kon
en
trolle
w
k
k

onnenurther
Extended


The
Mathematicians
to
pro
kno
v
algorithms
e
pro

the
theorems
T
of
plan
a
the

,
mathematical

domain
ys
b
ultiple
y
algorithms
exi-
algorithm
bly
ws

rules,
bining

sev
of
eral
This
global
tiation
and

lo
kno

e,
problem

solving
w
strategies.
motiv
In
the

for
trast,
to
most
algorithm
of
the
to
individual
da
with
y's

automated
pro
theorem
domain
pro
ematical)
ving
the
systems
to
use
sub
one
en-
or
application,
few
king.
strategies
an
and
algorithm
t
of
yp-
ded
ically
erse.
their
t


trol
dieren
is
king
hard-co

ded
ariables.
in
of
to
the
the
of
systems
ork
algorithms.
of
This
that
w
ving
as
so
also
of
true
and
for
theorem

exible
mega
pro
's
out
previous
stated
pro
trol
of
so-called
planner,
reason
whic

h
,

of
bined
further
the
kno
application
osition
of
monolithic
planning
allo
op
and
erators,
of
the
onen
instan
in
tiation
t
of
op
v
ariable
ariables,

and
ec

strategy

tiation
king
a
in
determines
a
eha
pre-dened
algorithm.
w
en-
a
to
y
e
.

Moreo

v
hniques
er,
e
the
problems.
functionalities
describ
of
w
these

sub
dieren

ys
onen
ob
ts
tiate
w
the
ere
for
v
with
ery
w

osition
The
monolithic
hard-co
pro
ded
frame-

op
bination
in
of
kinds
op
their
erations

with
theorem


functionalities
in
prohibited
are
the
the
use
steps,
of

mathematical
algorithm
kno
of
wledge
v
of
enable

bination
pro
m
of
planning

meta-reasoning
and
applicable
their
ativ


bina-

tion.
wledge
As
in
a

result,
h
the
out
planner
plan
failed
far,
on

problems
the
for
the
whic

h
de
more
(math-
exibilit
domain
y
wledge.
and

kno
of
wledge
previous
is
pro
needed
planner
in
ws
the
extend
pro
generalize
of
functionalities
planning
its
pro


ts.
These
results
observ
indep
ations
den
led
parameterized
us
for
to
erator
dev
v
elop
instan
pro
and
of

planning
T
with
hnically
m
a
ultiple
is
strategies,
instan
whic
of
h
h
w
parameterized
e
and
in
a
tro
b

vior
in
the
this
The
thesis.
wledge
The

main
in
idea
strategies
is
b
to
div

Strategies
ose
describ
the
for
previous
dieren
monolithic

pro
to
of
v
planning
a
pro
of

Strategies
and
also
to
e
replace
t
it
a
b
of
y

separate
or
but
t

a
orating
of
op
mathematical
erations,

so-called
instan
strategies,
v
whic
Although
h
initial

ation
realize
pro
dieren
planning
t
m
plan
strategies
renemen
as
ts

and
of
mo
previous
di