Scribd
Upload
27 views29 pages

Chapter 7

This document discusses properties of random graphs with n nodes and edge probability p. It can be summarized in 3 points: 1. The average number of edges in the graph is p*n(n-1)/2 and the average degree of a node is p(n-1). As n increases, p must decrease for the graph to remain sparse. 2. There is a phase transition in the graph structure at the critical value p=1/n. For p<1/n, the largest component is small. For p>1/n, a giant component emerges containing a finite fraction of nodes. 3. As n increases, the node degree distribution approaches a Poisson distribution with mean

Uploaded by

mahmoud sersawy
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
27 views29 pages

Chapter 7

This document discusses properties of random graphs with n nodes and edge probability p. It can be summarized in 3 points: 1. The average number of edges in the graph is p*n(n-1)/2 and the average degree of a node is p(n-1). As n increases, p must decrease for the graph to remain sparse. 2. There is a phase transition in the graph structure at the critical value p=1/n. For p<1/n, the largest component is small. For p>1/n, a giant component emerges containing a finite fraction of nodes. 3. As n increases, the node degree distribution approaches a Poisson distribution with mean

Uploaded by

mahmoud sersawy
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 29

Gin p It is random

a
graph
with n nodes and p the probability
that pairs of nodes
each are connected
The number of edges m in this model
IVI n El spited
d f
the number the number of
of nodes edges

Since we don't know the number of


edges we can find the average
number of edges will be

m
2 P
IIP
30 the average meandegree in
a network with exactly m edges is
so the mean degree
2yd
u
fan 241 MYIP
In 1 P Imp D C M IIP
so the probability that a particular node
v has degree K will follow the binomial
distribution

i k
pi
P degios ie
É p pi ms

In many cases we are interested in the

properties of large networks so that


n can be assumed to be large
Furthermore almost all real world networks
are sparse meaning that the only a tiny
fraction of the possible edges are
y
actually present and the average degree X
is much less than n More formally
a sparse network is one in which the

average degree increases slower than n

a s n becomes large In which Case

eg implies that p will become

vanishingly small
From

t
in ppg n t t

n t k
FThy X using
taylor series
and the equalities become exact as n n

with K fied Taking e for both sides


t
i pi E
Also for large n we have

Italia
D Pldeglutte c pké
é é x
e F Ten
in the limit of large n

This is the passion dist


AS X grows this dist in turn to

graph
i
Ck
p o empty o

pal complete graph Ck no l

ng largest connected component


s NG
Phase Transition
Let a fraction of nodes that do not belong
to Gcc The probability that a node
doesn't belong to Gcc

U
heung PCR o P CK Dut p K 2 UH

É Perak
s
E EE u

e e Eca
Let s fraction of nodes belonging to Gcc
Size of Gcc
cu l
U et
sit u
III's S l É

X ph Lies
when X A S l

when X o s o

let y i Ets

Mean degree I

non Zero solution exists when Cats o

Take the derivative of both sides


of
l é I

xé I
at s o 7 1 Lk pn I
Pent
In other words the random graph can have
a gaint component only if X 1
if X C I p no solution
Graph Catnip for n a critical value
Pet
Subcritical regime Pep a I

there is no components with more than


0 Inn nodes largest component is a tree
Critical point P Py Ck the largest component

has 0 n's nodes


supercritical regime p P X 51
gigantic component has all
OCCP Pc n nodes
Connected regime Ps In mln 4K Inn
gigantic component has all 0 n nodes
critical value In P n

I on average one
neighbor for a
node
pPit p

You might also like