The NLC width is a term from graph theory and assigns a natural number to each undirected graph . Certain difficult problems such as MAX-CUT or the Hamilton cycle problem can be solved in polynomial time on graphs with a limited NLC width .
definition
The concept of the NLC width was introduced by Wanke in 1994. The term k-marked graph is important for the definition of the NLC width :
k-marked graph
For one be
k
∈
N
{\ displaystyle \, k \ in \ mathbb {N}}
[
k
]
: =
{
1
,
...
,
k
}
{\ displaystyle \ lbrack k \ rbrack \,: = \ lbrace 1, \ ldots, k \ rbrace}
A k-marked graph is a graph whose nodes are marked with a marker image
G
=
(
V
G
,
E.
G
,
l
a
b
G
)
{\ displaystyle \ G = (V_ {G}, E_ {G}, lab_ {G}) \}
(
V
G
,
E.
G
)
{\ displaystyle \ (V_ {G}, E_ {G}) \}
l
a
b
G
:
V
G
→
[
k
]
{\ displaystyle \, lab_ {G}: V_ {G} \ rightarrow \ lbrack k \ rbrack}
A graph with exactly one node marked with is denoted by
i
∈
[
k
]
{\ displaystyle i \ in \ lbrack k \ rbrack}
∙
i
{\ displaystyle \ bullet _ {i}}
definition
The NLC width of a k-marked graph is the smallest natural number so that it belongs to the graph class .
G
{\ displaystyle G}
k
,
{\ displaystyle k,}
G
{\ displaystyle G}
N
L.
C.
k
{\ displaystyle NLC_ {k}}
It is defined recursively as follows:
N
L.
C.
k
{\ displaystyle NLC_ {k}}
The -marked graph is for an in
k
{\ displaystyle k}
∙
i
{\ displaystyle \ bullet _ {i}}
i
∈
[
k
]
{\ displaystyle i \ in \ lbrack k \ rbrack}
N
L.
C.
k
{\ displaystyle NLC_ {k}}
Be and in . Furthermore, let and be vertically disjoint and a relation. Then there is also the -marked graph
G
=
(
V
G
,
E.
G
,
l
a
b
G
)
{\ displaystyle G = (V_ {G}, E_ {G}, lab_ {G})}
J
=
(
V
J
,
E.
J
,
l
a
b
J
)
{\ displaystyle J = (V_ {J}, E_ {J}, lab_ {J})}
N
L.
C.
k
{\ displaystyle NLC_ {k}}
G
{\ displaystyle G}
J
{\ displaystyle J}
S.
⊆
[
k
]
2
{\ displaystyle S \ subseteq \ lbrack k \ rbrack ^ {2}}
k
{\ displaystyle k}
G
×
S.
J
=
(
V
′
,
E.
′
,
l
a
b
′
)
{\ displaystyle G \ times _ {S} J = (V ', E', lab ')}
in , with
N
L.
C.
k
{\ displaystyle NLC_ {k}}
V
′
=
V
G
∪
V
J
{\ displaystyle V '= V_ {G} \ cup V_ {J}}
,
E.
′
=
E.
G
∪
E.
J
∪
{
{
u
,
v
}
|
u
∈
V
G
,
v
∈
V
J
,
(
l
a
b
G
(
u
)
,
l
a
b
J
(
v
)
)
∈
S.
}
{\ displaystyle E '= E_ {G} \ cup E_ {J} \ cup \ lbrace \ lbrace u, v \ rbrace | u \ in V_ {G}, v \ in V_ {J}, (lab_ {G} ( u), lab_ {J} (v)) \ in S \ rbrace}
and
l
a
b
′
(
u
)
=
{
l
a
b
G
(
u
)
,
if
u
∈
V
G
l
a
b
J
(
u
)
,
if
u
∈
V
J
{\ displaystyle lab '(u) = {\ begin {cases} lab_ {G} (u), & {\ text {falls}} u \ in V_ {G} \\ lab_ {J} (u), & { \ text {falls}} u \ in V_ {J} \ end {cases}}}
for everyone .
u
∈
V
′
{\ displaystyle u \ in V '}
Let be a -marked graph and a total function. Then there is also the -marked graph
G
=
(
V
G
,
E.
G
,
l
a
b
G
)
∈
N
L.
C.
k
{\ displaystyle G = (V_ {G}, E_ {G}, lab_ {G}) \ in NLC_ {k}}
k
{\ displaystyle k}
R.
:
[
k
]
→
[
k
]
{\ displaystyle R: \ lbrack k \ rbrack \ rightarrow \ lbrack k \ rbrack}
k
{\ displaystyle k}
∘
R.
(
G
)
=
(
V
G
,
E.
G
,
l
a
b
′
)
{\ displaystyle \ circ _ {R} (G) = (V_ {G}, E_ {G}, lab ')}
in with .
N
L.
C.
k
{\ displaystyle NLC_ {k}}
l
a
b
′
(
u
)
=
R.
(
l
a
b
(
u
)
)
∀
u
∈
V
G
{\ displaystyle lab '(u) = R (lab (u)) \ qquad \ forall u \ in V_ {G}}
example
The following table demonstrates the construction of the "House of Nicholas" graph using the operations defined above:
NLC operation
Representation of the graph
G
1
=
∙
1
×
{
(
1
,
2
)
}
∙
2
{\ displaystyle G_ {1} = \ bullet _ {1} \ times _ {\ lbrace (1,2) \ rbrace} \ bullet _ {2}}
G
2
=
∙
3
×
{
(
3
,
4th
)
}
∙
4th
{\ displaystyle G_ {2} = \ bullet _ {3} \ times _ {\ lbrace (3,4) \ rbrace} \ bullet _ {4}}
G
3
=
G
1
×
{
(
1
,
3
)
,
(
1
,
4th
)
,
(
2
,
3
)
,
(
2
,
4th
)
}
G
2
{\ displaystyle G_ {3} = G_ {1} \ times _ {\ lbrace (1,3), (1,4), (2,3), (2,4) \ rbrace} G_ {2}}
G
4th
=
∘
{
(
1
,
1
)
,
(
2
,
1
)
,
(
3
,
3
)
,
(
4th
,
3
)
}
(
G
3
)
{\ displaystyle G_ {4} = \ circ _ {\ lbrace (1.1), (2.1), (3.3), (4.3) \ rbrace} (G_ {3})}
G
N
i
k
O
l
a
u
s
=
G
4th
×
{
(
3
,
2
)
}
∙
2
{\ displaystyle G_ {Nikolaus} = G_ {4} \ times _ {\ lbrace (3,2) \ rbrace} \ bullet _ {2}}
It is therefore true . still has an NLC expanse of 1 because is a co-graph .
G
N
i
k
O
l
a
u
s
∈
N
L.
C.
4th
{\ displaystyle G_ {Nikolaus} \ in NLC_ {4}}
G
N
i
k
O
l
a
u
s
{\ displaystyle G_ {Nikolaus}}
G
N
i
k
O
l
a
u
s
{\ displaystyle G_ {Nikolaus}}
NLC range of special graph classes
The NLC widths of the following graph classes can be estimated upwards as follows:
Each co-graph has an NLC width of 1
Trees have an NLC width of 3 or less
Circles have an NLC width of 4 or less
Relation to the size of the clique
For every undirected graph with NLC-width and clique-width the following applies:
G
{\ displaystyle G}
n
l
c
w
(
G
)
{\ displaystyle nlcw (G)}
c
w
(
G
)
{\ displaystyle cw (G)}
n
l
c
w
(
G
)
≤
c
w
(
G
)
≤
2
⋅
n
l
c
w
(
G
)
.
{\ displaystyle nlcw (G) \ leq cw (G) \ leq 2 \ cdot nlcw (G).}
literature
Frank Gurski, Irene Rothe, Jörg Rothe, Egon Wanke: Exact algorithms for difficult graph problems , Springer-Verlag, Berlin Heidelberg, 2010, ISBN 978-3-642-04499-1
Individual evidence
↑ Egon Wanke: k -NLC Graphs and Polynomial Algorithms, Discrete Applied Mathematics, 54: 251-266, 1994
<img src="https://de.wikipedia.org/wiki/Special:CentralAutoLogin/start?type=1x1" alt="" title="" width="1" height="1" style="border: none; position: absolute;">