<s>
In	O
artificial	B-Application
intelligence	I-Application
,	O
hierarchical	B-Application
task	I-Application
network	I-Application
(	O
HTN	O
)	O
planning	O
is	O
an	O
approach	O
to	O
automated	B-Application
planning	I-Application
in	O
which	O
the	O
dependency	O
among	O
actions	O
can	O
be	O
given	O
in	O
the	O
form	O
of	O
hierarchically	O
structured	O
networks	O
.	O
</s>
<s>
primitive	O
(	O
initial	O
state	O
)	O
tasks	O
,	O
which	O
roughly	O
correspond	O
to	O
the	O
actions	O
of	O
STRIPS	B-Application
;	O
</s>
<s>
goal	O
tasks	O
(	O
goal	O
state	O
)	O
,	O
which	O
roughly	O
corresponds	O
to	O
the	O
goals	O
of	O
STRIPS	B-Application
,	O
but	O
are	O
more	O
general	O
.	O
</s>
<s>
One	O
particular	O
formalism	O
for	O
representing	O
hierarchical	B-Application
task	I-Application
networks	I-Application
that	O
has	O
been	O
fairly	O
widely	O
used	O
is	O
TAEMS	B-Application
.	O
</s>
<s>
Nonlin	O
,	O
one	O
of	O
the	O
first	O
HTN	B-Application
planning	I-Application
systems	O
.	O
</s>
<s>
UMCP	O
,	O
the	O
first	O
provably	O
sound	O
and	O
complete	O
HTN	B-Application
planning	I-Application
systems	O
.	O
</s>
<s>
PANDA	O
,	O
a	O
system	O
designed	O
for	O
hybrid	O
planning	O
,	O
an	O
extension	O
of	O
HTN	B-Application
planning	I-Application
developed	O
at	O
Ulm	O
University	O
,	O
Germany	O
.	O
</s>
<s>
HTNPlan-P	O
,	O
preference-based	B-Application
HTN	B-Application
planning	I-Application
.	O
</s>
<s>
HTN	B-Application
planning	I-Application
is	O
strictly	O
more	O
expressive	O
than	O
STRIPS	B-Application
,	O
to	O
the	O
point	O
of	O
being	O
undecidable	O
in	O
the	O
general	O
case	O
.	O
</s>
<s>
However	O
,	O
many	O
syntactic	O
restrictions	O
of	O
HTN	B-Application
planning	I-Application
are	O
decidable	O
,	O
with	O
known	O
complexities	O
ranging	O
from	O
NP-complete	O
to	O
2-EXPSPACE-complete	O
,	O
and	O
some	O
HTN	O
problems	O
can	O
be	O
efficiently	O
compiled	O
into	O
PDDL	B-Application
,	O
a	O
STRIPS-like	O
language	O
.	O
</s>
