Information Technology Reference
In-Depth Information
Proof.
The in-degree sum formula states that
v
∈
G
deg
−
(
v
)=
m
.
From this we have that the parity of
O
d
,
m
equals the parity of
m
. Note that
d
−
E
d
,
m
=
O
d
,
m
. Hence the parity of
m
equals the parity of
d
−
E
d
,
m
and the lemma follows.
For any real number
r
, we denote
r
+
=
max
{
r
,
0
}
.
Lemma 7.
For every oriented graph on d vertices and m edges,
Z
d
+
v
(
deg
−
(
v
)
−
1
)
+
=
m
−
d
.
Proof.
We use again the in-degree sum formula,
−
∑
v
deg
−
(
deg
−
v
)=
m
. Thus
∑
v
(
deg
−
(
)
+
=
(
v
)
−
1
)=
m
−
d
and so
−
Z
d
+
∑
v
(
v
)
−
1
m
−
d
, since the sum contributes
a
−
1 for every vertex of in-degree zero.
We denote by Med
(
Y
)
the median of the random variable
Y
.
d
−
2
(
E
d
)
≥
Lemma 8.
If Med
,for
any
graph on d vertices and d edges, then Theo-
2
rem
2
holds true.
deg
−
(
)
+
,since
=
E
d
−
Z
d
, then Lemma
7
gives that
Z
d
=
∑
v
(
)
−
Proof.
Let
Y
d
:
v
1
=
m
d
. Note that
∑
v
(
deg
−
(
v
)
−
1
)
+
≥
∑
deg
−
(
)
+
≥
∑
(
v
)
−
1
1
≥
Y
d
.
v
:deg
−
(
v
:deg
−
(
{
v
)
≥
2
}
{
v
)
≥
2
}
1
Since
Y
d
+
Z
d
=
E
d
, it follows that
Z
d
≥
2
E
d
.Now
X
d
+
Z
d
=
d
so that
X
d
=
−
Z
d
≤
d
1
2
E
d
and Med
d
−
2
3
d
+
2
d
−
(
X
d
)
≤
d
−
=
.
4
4
2
2
. To prove this, we first compute the
distribution of the number of even in-degree vertices in the case of a
connected
graph on
d
vertices and
m
d
−
So it remains to prove that Med
(
E
d
)
≥
≥
−
1 edges. We then extend this computation to the
general case by considering the connected components of the graph.
We denote by Bin
d
(
s
,
p
)
a Binomially distributed random variable of parameters
1
2
s
and
p
. In case
p
. The parity of the in-degree of each
particular vertex is related to the parity of the Binomial distribution for which the
following is well known.
=
we just write Bin
(
s
)
Lemma 9.
Suppose that X
s
:
=
Bin
(
s
)
mod
2
.ThenX
s
is a Bin
(
1
)
random variable
regardless of s.
Proof.
The proof is by induction on
s
.When
s
=
1 the conclusion is true. Suppose
that it is true for all integers up to
s
−
1 and consider
X
s
. Observe that
X
s
∼
X
s
−
1
+
Bin
(
1
)
,
mod 2. The induction hypothesis gives that
X
s
−
1
+
Bin
(
1
)
equals Bin
(
1
)+
Bin
(
1
)
mod 2,for two independent Bin
(
1
)
random variables which finishes the proof
of the lemma.
Search WWH ::
Custom Search