You are on page 1of 12

Proposition ③

Cycles and cut


Edges
to
Det E) Adding an
edge a
graph
two
combines the vertex sets of
.

of obtained comp
G-
subgraph 6
.

0 :

its into leaves the vertex sets of


deleting
b or
a card one

incident edges ) from G the components unchanged .

obtained
G- e
subgraph of 6
if if
'

at
µ
:
and
ciy An edge is
edge
a
only
by deleting e from G
it to
it is a cut vertex
-
of G belongs no
cycle
if # components (G- v1 Proof Li ) ✓

> # comp (6)


cut
.

Cio ) The following are


equivalent ,

e is a
-

of 6edge if
(1) ur is a cut
edge of 6

(4)
-

# comps G-e) > # comp (G)


(2) u and v are in distinct comport
of G- no
8 nottvpath
TIE
0 (3) 6- ur has

④ •
u v
e :

:
cut
edge
cut vertices
6 has

⑦#
no
cycle containing ur

,
Corollary with vertices and
Def Nontrivial graph ②
a)
Every graph n = at least one
edge
components
.

at has least k
edge
R n -

-
>

connected vertex
a) Every
n
graph Proposition nontrivial
graphics
Every
-

has at least edges which


M-1 ,
has at least two vertices that It
is sharp for paths not cut vertices Assume God
0-0-0-0 -0
are -

{
proof :

proof consider a maximal


then vertices
,
a) start with only ( non extendible) path P in G.
→ n components Let be an endpoint of P
u
each edge added decreases p
the number of components o o_0
ow
• - o - o -
- - .
-

at most
by is a

cannot be extended
→ zn -
R components since P ,

Nlu) EP
after k
edges disconnected because
so G- u is not
So need to add n-2
(2) we
one can use P to
go from any
vertices to get 71 comp to another
neighbor of u
.

.

The
Pr -

Y Lemurs :
A
graph with 86) 2

u⑧←=y€h¥Ép
.

has

• -
a
cycle
Proof , Let P be a maximal
path of 6 and u an endpoint .

P1 since
d§u)zz then u has

another neighborCliffe rent


from the one next in P
if there is a
path in G :

If
xu%v→u→wnE >
y u

-
P D

where v. WENCH then Def Adecomposition of graph 6 a

v. → u→w is a
family of subgraphs that

be replaced with the partition EG) An even graph


-

can .

where
swbpath of P between wand w is a
graph every vertex
vii> w degree is even .

ff

✗ Theorem :
in> y Every even
graph
decomposes into cycles
induction tbjosconjectvre.TV ④
-

Proof By on m -

# edge >

m= 0
✓ Every
vertex
n -

graph
even

decomposes into £4k ) cycles


m > o : let H be a nontrivial
component of G

SCH ) zz
Galibi conjecture
connection vertex graph
cycle
3-
'

C in H Every -

decomposes into £ TMI


Deleting the
edges of C fromH
' paths
leaves even
an
graph G
Theorem ( Lovasz) Every n vertex
-

"

HEY! :L:*
into £ Lnlzj
graph decomposes
.

BY I. H G
'
has a
deoomp
paths and cycles
.

--
Add C to obtain a

Drop A K regular graph Gg


decomposition for 6
-

(Kyra)
"

decomposes
"
☐ into Stars

EI if and only if Gibiprtte


- T
Assante there decomposition
Kun
%j
is a

"
into stars -

. .

1 3
Let ✗ be the set of centers
z

of these stars and


proof .

be the other
Y
⇐/ Assume 6
regular bipartite
¥
b-
vertices .

with birsartitwn Xie "


tripartition
.

×, Y is a

Make a star with each REX ☐

as center and k neighbors


in Y EULERIAN CIRCUITS

±÷E⇐⇐:
Konigsberg Bridge Problem

river
r

Sendero ⑥
Is it possible to take a stroll circuit 1 Closed tail

crossing each
bridge exactly Eulerian trail : Trail that
once ? "
traverse> t all edge>

"

04
Eulerian circuit Closed eulerian
Abstraction :
2.2 .

2¥⑧⑥#
trail

JD
.

Theorem A multi has


:

graph 6

6 4 an Eulerian circuit
ifaoonly
if it has at most one nontrivial
component and all vertex
degrees
are even .

Def Multi
graph Proof :
) An Eulerian circuit
loops
µ
.

Allows in all vertices


has even
degree
and multi
edges .
and is connected

Def Trail [sendero ] : walk ⇐) 6 has a decomposition


that traverses each into
cycles . This is also a

edge once .
decomposition into circuits .

Their 6
decomposes into@

Consider a
decomposition into circuits trails and no fewer
smallest
.

S of size .

Claim : Is 1--1
-

Ll -

Algorithm for finding


.

If 1st > 1 sndno pair of circuits


an

Euler tour
have a common vertex ,
then

u¥*?÷¥☒¥É
6 would have more than I

ants
component .
so - @

vertex

must share a . and
into
they could be
joined • _€ •
contradiction Greedy fails
"

L circuit .
A "
.

shop ,din¥s

For multi
graphs Solutions Give priority to

dt-cv-d.tv/foz-G .

non cut
edges in the
when
remaining graph
-

connected multigsph
corollary G
selecting the next edge
z④
.

with odd vertices


=
degree Bondy & Murty (mien)
0 a- last vertex of ✗ ⑧
- visited .

0
3- 2- C-
Nfgtk)
not a
Nz was
ceitedge
when eo
Ny was
=

chosen ☒
Suppose the
algorithm finishes
without traversing all
edges
Let
Ff . . . find F

Wf - .
final w

E ( Ff ) = E- (G) -
E (Wf )
EV vertices v with
✗ :

dhgfgco ) > 0

'

Obs_ G=Cv,E) °↳
✗ EV

do IEscxj.lt/zg(xy
i t 0*0 V4

E(G[×] ) edges out

of the 4- ✗
edges if degrees
2€2
in ✗ are even
induced
subgraph
✗ then contradiction in
by
[ dew )=2 / EdD / +12801
*
even even L .

Vx

TREES AND DISTANCE ⑨
Det :

Acyclic graphs graph with ☒☒☒→⇐ spanning


tree
no
cycles
Tree : connected
acyclic graph Properties of trees
Det A leaf of
Ig
tree
Forest whose components a
:
graph .

are trees is a vertex of degree 1

tree
spanning :
spanning Pop ! "
Every
nontrivial tree
that
)
recubriobr .

subgraph has at least 2 leaves


.
is also tree

÷¥÷¥¥÷÷*
.

leaf from tree


(2) Deleting a a

tree
yields a

proof a) The endpoints of a

maximal path are leaves


and connected
of G- it is
acyclic ☐
Ikram proof ①
6
nierxgraph
: .
. .

The
following are
equivalent : A-
B) Poriuducaon
sobre n
A) G is connected and
acyclic
.

B) G is connected and
c-
has
tree )

n -2 ) n=1
✓ •

edgett n > I Exisleomhojz


-

G has and is
c) n -1
edges UEUT)
acyclic T1=T- u
-

airbed

D) fu OEVCG) , ,
there is
unique
a
conn-1 vest
V.v path
.

zanstas
-

n -

connected '
E) 6 is
minimally : HI en T

(F)
CG is connected but teeth ns.ansbs
G- e is disconnected )
out
maximally acyclic
.

6 is
6
16 is
acyclic but fuv¢Ek ) B c) es eonexo
y
has
G+Hy☒ a cycle ) I n-kens.to
~2

6 tiene
supongoos que
on ciob C.

podemos remover V92

arista de C sin de> corrector 6

ites nos hasta Sblenergstoaadico


< n-1 aristo

control cairn
,

A B
pogue
-

You might also like