Sei sulla pagina 1di 47

Generalized Firefighting on the 2

dimensional infinite grid


Kah Loon Ng
DIMACS
Centrality measures in social
networks
and
Containing fires in infinite grids L
d

Fire starts at only one vertex:
d= 1: Trivial.
d = 2: Impossible to contain the fire with 1 firefighter per
time step
Containing fires in infinite grids L
d

d = 2: Two firefighters per time step needed to contain the fire.
8 time steps
18 burnt
vertices
Containing fires in infinite grids L
2

We assume that the number of firefighters available for
deployment is given by a function that is periodic. ) (t f
Justification for considering periodic functions:
1 = t 2 = t 3 = t 4 = t 5 = t
6 = t
1 1 2 1 1 2
Containing fires in infinite grids L
2

Given a periodic function , let f
f
p
= period of
f

=
=
f
p
i
f
i f N
1
) (
f
f
f
p
N
R =
We also write
] , , [ 3 2 1 = f
For example,

=
3
2
1
) (x f
if
if
if
1 x
2 x
0 x mod 3
mod 3
mod 3
,...) , , , , , ( 3 2 1 3 2 1
6 =
f
N 2 =
f
R
Containing fires in infinite grids L
2

Given positive integers and , we define a periodic function n
k
] , ,... , , [
,
k g
n
k n

1
1 1 1 1

=
Given two periodic functions and with periods and
respectively, we say
f g
f
p
g
p
N e n g f - |
.
|

\
|

s
= =
n
i
n
i
i g i f
1 1
) ( ) (
) , (
g f
p p lcm ne
|
.
|

\
|

s
= =
n
i
n
i
i g i f
1 1
) ( ) (
Containing fires in infinite grids L
2

If is a periodic function and is any positive integer, define
the translate function by
f i
i
f
+
) ( ) ( i x f x f
i
+ =
+
For example, if
,...) , , , , , , , , , ( ] , , , , [ 3 2 2 1 1 3 2 2 1 1 3 2 2 1 1 = = f
and
3 = i
then
,...) , , , , , , , , , ( ] , , , , [ 2 1 1 3 2 2 1 1 3 2 2 1 1 3 2
3
= =
+
f
Note that
f f
p p
i
=
+
f f
N N
i
=
+
f f
R R
i
=
+
Containing fires in infinite grids L
2

5 1. >
n
M
R
Note that is chosen to be the smallest integer such
that .
o
Define to be the periodic function of period
n
D
1 2 + n
] ,..., , , ,..., , [
_

1
2 2 2 1 1 1
+
=
n
n
n
D
Given a positive integer , let be the function with period
defined by
n
] , ,... , , [ o

1
1 1 1 1

=
n
n
M
n
M
n

2
2
+ =
n
o
where
Containing fires in infinite grids L
2

Main Theorem:
If the number of firefighters available for deployment per time step
can be represented by a periodic function such , then
any fire that breaks out at a single vertex in can be contained
after some finite time .
5 1. >
f
R f
2
L
t
Lemma 0:
Suppose and are two periodic functions such that . If
any fire breaking out at a single vertex in can be contained
using , then it can also be contained using .
f g
g f -
2
L
f
g
Containing fires in infinite grids L
2

Proof of Lemma 0:
Let . ) , (
g f
p p lcm s =
) ( ) ( i g i f s s i s s 1 Case 1: If for all , we are done.
) ( ) (
* *
k g k f x =
Case 2: If for some . Let be the smallest
such .
) ( ) ( k g k f >
2 > k
*
k
k
Since , we have . g f -
>

=
1
1
*
) ( ) (
k
i
x i f i g
Time 1 to : Deploy firefighters as in , and
accumulate at least spare firefighters.
1
*
k
f
x
Containing fires in infinite grids L
2

Outline proof of Main Theorem:
Given any periodic function
with
f
5 1. >
f
R
~
Rearrange to get
that is non decreasing in its
period.
f ' f
] , , , , , [ 2 1 4 1 3 2
~
] , , , , , [ 4 3 2 2 1 1
] , , , , , [ 4 3 2 2 1 1 ~ ] , , , , , [ 8 1 1 1 1 1
~
' f
) ,( 1 +
f f f
p N p
g
Lemma 1:
~
) ,( 1 +
f f f
p N p
g
f
p
M
] , , , , , [ 8 1 1 1 1 1 ~ ] , , , , , [ 5 1 1 1 1 1
Containing fires in infinite grids L
2

Outline proof of Main Theorem:
) ,( 1 +
f f f
p N p
g
~
f
p
M
] , , , , , [ 8 1 1 1 1 1 ~ ] , , , , , [ 5 1 1 1 1 1
~
f
p
M 2
) (
f
p
D
] , , , , , [ 5 1 1 1 1 1 ~ ] ,..., , , ,..., , [
_ _
37 36
2 2 2 1 1 1
Lemma 4:
1,1,1,1,5, 1,1,1,1,5, 1,1,1,1,5, 1,1,1,1,5, 1,1,1,1,5,
35
1,1,1,1,1,.1,1,2,2,.
36
Containing fires in infinite grids L
2

Outline proof of Main Theorem:
f
p
M ~ 2
) (
f
p
D
] , , , , , [ 5 1 1 1 1 1 ~ ] ,..., , , ,..., , [
_ _
37 36
2 2 2 1 1 1
Lemma 4:
Lemma 2: If is a periodic function that is non decreasing
over its period, then for any positive integer , .
f
i
f f
i
~
+
1, 1, 2, 2, 3, 1, 1, 2, 2, 3,
f
2 +
f
2, 2, 3, 1, 1, 2, 2, 3, 1, 1,
Gain Loss
Containing fires in infinite grids L
2

Outline proof of Main Theorem:
f
p
M ~ 2
) (
f
p
D
] , , , , , [ 5 1 1 1 1 1 ~ ] ,..., , , ,..., , [
_ _
37 36
2 2 2 1 1 1
Lemma 4:
Lemma 3: If is a periodic function that is non decreasing
over its period, then for any periodic function with
f
g
f g
p p >

s
= =
k
i
k
i
i f i g
1 1
) ( ) (
for ,
g
p k s s 1
In other words,
g f ~

s
= =
k
i
k
i
i f i g
1 1
) ( ) (
for , ) , (
g f
p p lcm k s s 1
satisfying we have .
1,1,.,1, 2,2,,2,2,1,1,..
1,1,1,1,1,5,...1,1,1,1,1,5, 1,1,1,1,1,5,..1,5,1,1,1,..
36 37
Containing fires in infinite grids L
2

The strategy using ] ,..., , , ,..., , [

1
2 2 2 1 1 1
+
=
n n
n
D
End of phase 1
End of phase 2
End of phase 3
Advance firefighters
Retreat firefighters
Containing fires in infinite grids L
2

The strategy using :
n
D
Completing phase 1:
1
3
5
7
9
11
2
4
6
7 8
8 9 10
10
11 12
12
] , , , , , , , , , , , , [ 2 2 2 2 2 2 2 1 1 1 1 1 1
6
= D
Containing fires in infinite grids L
2

The strategy using :
n
D
Delayed response Phase 1 can still be completed.
Containing fires in infinite grids L
2

The strategy using :
n
D
Completing phase 2 after phase 1 has been completed:
(0,0)
(0,0)
Containing fires in infinite grids L
2

The strategy using :
n
D
Completing phase 3 after phase 2 has been completed:
(0,0)
(0,0)
Containing fires in infinite grids L
2

The strategy using :
n
D
Finishing the job:
1
Active
vertices
Retreat
firefighters
(0,0)
2
4
3
(0,0)
Containing fires in infinite grids L
2

A few points to note:
Our theorem says nothing about periodic functions where .
In fact, we can easily construct a function where and
still is sufficient to contain the outbreak.
f
5 1. s
f
R
g
c + =1
g
R
g
(0,0)
Containing fires in infinite grids L
2

A few points to note:
Our theorem says nothing about periodic functions where .
In fact, we can easily construct a function where and
still is sufficient to contain the outbreak.
f
5 1. s
f
R
g
c + =1
g
R
g
Obviously, is sufficient to contain the outbreak but
we do not consider such (cheating! ) situations.
] ,......, , , , [ 1 1 1 1 4
The exact time required to contain the outbreak (and thus the
number of burnt vertices) can be explicitly computed as a
function of . However, our strategy does not guarantee that the
number of burnt vertices is minimized.
n
Centrality Measures in Graphs (or Social Networks)
Centrality = Importance = Prominence?
3 types of centrality indices:
degree closeness
betweeness
Centrality Measures in Graphs (or Social Networks)
Centrality indices can be computed for each vertex or a
group of vertices.
Majority of centrality concepts are based on non-directed
and dichotomous relations
However, for some specific purposes (for example,
measuring prestige) directed graphs or valued relations might
need to be used.
Centrality Measures in Graphs (or Social Networks)
) (
i a
n C
= centrality index for vertex under measure
i
n a
a = D, degree measure
a = C, closeness measure
a = B, betweeness measure


=
=
=
n
i
i a a
n
i
i a a
a
n C n C
n C n C
C
1
1
)) ( ) ( ( max
)) ( ) ( (
*
*
Centrality Measures in Graphs (or Social Networks)
) (
i D
n C (degree centrality)
) (
i D
n C = deg = degree of vertex ) (
i
n
i
n
= deg ) ( '
i D
n C ) (
i
n
1
1
n


=
=
=
n
i
i D D
n
i
i D D
D
n C n C
n C n C
C
1
1
)) ( ) ( ( max
)) ( ) ( (
*
*
) )( (
)) ( ) ( (
*
2 1
1


=
=
n n
n C n C
n
i
i D D
n
C n C
S
n
i
D
i D
D


=
=1
2
2
) ) ( (
Centrality Measures in Graphs (or Social Networks)
) (
i C
n C (closeness centrality)
1
1

= )) , ( ( ) (
j
n
j
i i C
n n d n C


=
=
=
n
i
i C C
n
i
i C C
C
n C n C
n C n C
C
1
1
)) ( ' ) ( ' ( max
)) ( ' ) ( ' (
*
*
) /( ) )( (
)) ( ) ( (
*
3 2 1 2
1


=
=
n n n
n C n C
n
i
i C C
) ( ) (
) , (
) ( '
i C
n
j
j i
i C
n C n
n n d
n
n C 1
1
1
=


=
=
Centrality Measures in Graphs (or Social Networks)
) (
i C
n C (closeness centrality)
Closeness centrality measures are related to:
Jordan centers of a graph (subset of with
the smallest eccentricities)
) (G V
Centroid of a graph (subset of with the
smallest weight)
) (G V
One shortcoming of closeness measures is that it cannot be
used for disconnected graphs.
Centrality Measures in Graphs (or Social Networks)
) (
i B
n C (betweeness centrality)
Let
jk
g
= number of distinct shortest paths between and j
k
= number of distinct shortest paths between and that
passes through
j
k
i
n
) (
i jk
n g

=
<k j
jk i jk i B
g n g n C / ) ( ) (
For each , maximum
value is 1.
k j <
|
|
.
|

\
|

=
<
2
1 n
g n g
n C
k j
jk i jk
i B
/ ) (
) ( '
1
1


=
=
n
n C n C
C
n
i
i B B
B
)) ( ' ) ( ' (
*
Centrality Measures in Graphs (or Social Networks)
) (
i B
n C (betweeness centrality)
We assume here that if some information (or disease) is
passed from to , each of the shortest paths is equally
likely to be chosen.
j k
If we sum over all , we obtain measures of
the pair-dependency of vertex on vertex . These values
can also be viewed as indices of how much gate keeping
does for .
jk i jk
g n g / ) (
k
j
i
i
n
j
n
seems to be an improvement over and
but there are still inadequacies.
) (
i B
n C ) (
i D
n C ) (
i C
n C
Centrality Measures in Graphs (or Social Networks)
MEDICI
TORNABUONI
RIDOLFI
GUADAGNI
ALBIZZI
LAMBERTESCHI
GINORI
BISCHERI
STROZZI
CASTELLANI
PERUZZI
BARBADORI
SALVATI
ACCIAIUOLI
PAZZI
Centrality Measures in Graphs (or Social Networks)
Acciaiuoli
Albizzi
Barbadori
Bischeri
Castellani
Ginori
Guadagni
Lamberteschi
Medici
Pazzi
Peruzzi
Ridolfi
Salvati
Strozzi
Torabuoni
) ( '
i D
n C ) ( '
i C
n C ) ( '
i B
n C
0.071
0.214
0.143
0.214
0.214
0.071
0.286
0.071
0.429
0.071
0.214
0.214
0.143
0.286
0.214
0.368
0.483
0.438
0.400
0.389
0.333
0.467
0.326
0.560
0.286
0.368
0.500
0.389
0.438
0.483
0.000
0.212
0.093
0.104
0.055
0.000
0.255
0.000
0.522
0.000
0.022
0.114
0.143
0.103
0.092
Centrality Measures in Graphs (or Social Networks)
Acciaiuoli
Albizzi
Barbadori
Bischeri
Castellani
Ginori
Guadagni
Lamberteschi
Medici
Pazzi
Peruzzi
Ridolfi
Salvati
Strozzi
Torabuoni
) ( '
i D
n C ) ( '
i C
n C ) ( '
i B
n C
0.071
0.214
0.143
0.214
0.214
0.071
0.286
0.071
0.429
0.071
0.214
0.214
0.143
0.286
0.214
0.368
0.483
0.438
0.400
0.389
0.333
0.467
0.326
0.560
0.286
0.368
0.500
0.389
0.438
0.483
0.000
0.212
0.093
0.104
0.055
0.000
0.255
0.000
0.522
0.000
0.022
0.114
0.143
0.103
0.092
Some observations
Strozzi family
has high degree
centrality but
low closeness
centrality.
Centrality Measures in Graphs (or Social Networks)
Acciaiuoli
Albizzi
Barbadori
Bischeri
Castellani
Ginori
Guadagni
Lamberteschi
Medici
Pazzi
Peruzzi
Ridolfi
Salvati
Strozzi
Torabuoni
) ( '
i D
n C ) ( '
i C
n C ) ( '
i B
n C
0.071
0.214
0.143
0.214
0.214
0.071
0.286
0.071
0.429
0.071
0.214
0.214
0.143
0.286
0.214
0.368
0.483
0.438
0.400
0.389
0.333
0.467
0.326
0.560
0.286
0.368
0.500
0.389
0.438
0.483
0.000
0.212
0.093
0.104
0.055
0.000
0.255
0.000
0.522
0.000
0.022
0.114
0.143
0.103
0.092
Some observations
Tornabouni
family has high
closeness
centrality but
low betweeness
centrality.
Centrality Measures in Graphs (or Social Networks)
Acciaiuoli
Albizzi
Barbadori
Bischeri
Castellani
Ginori
Guadagni
Lamberteschi
Medici
Pazzi
Peruzzi
Ridolfi
Salvati
Strozzi
Torabuoni
) ( '
i D
n C ) ( '
i C
n C ) ( '
i B
n C
0.071
0.214
0.143
0.214
0.214
0.071
0.286
0.071
0.429
0.071
0.214
0.214
0.143
0.286
0.214
0.368
0.483
0.438
0.400
0.389
0.333
0.467
0.326
0.560
0.286
0.368
0.500
0.389
0.438
0.483
0.000
0.212
0.093
0.104
0.055
0.000
0.255
0.000
0.522
0.000
0.022
0.114
0.143
0.103
0.092
Some observations
257 0. =
D
C
322 0. =
C
C
437 0. =
B
C
Centrality Measures in Graphs (or Social Networks)
Eigenvector centrality
The centrality of a vertex does not depend on the number of
vertices it is adjacent to, but also these vertices centrality.
Bonacich (1972) defines (eigenvector) centrality
as a positive scalar multiple of the sum of adjacent
centralities:

=
=
n
j
j E ij i E
n C a n C
1
) ( ) (
) (
i E
n C
In matrix form, if we have
T
n E E
n C n C C )) ( ),..., ( (
1
=
C AC =
(Perron-Frobenius) Since is nonnegative, there exists an
eigenvector of the maximal eigenvalue with only
nonnegative entries.
A
Centrality Measures in Graphs (or Social Networks)
Information centrality
Adopts the idea that information is passed along the network
along all possible paths, not necessary the shortest one.
Flow betweenes centrality
The flow betweeness of vertex is defined as the amount of
flow through vertex when the maximum flow is transmitted
from to , averaged over all and .
i
i
s t s t
Random walk betweenes centrality
The random walk betweeness of a vertex is the number of
times that a random walk starting at and ending at passes
through along the way.
i
s t
i
Centrality Measures in Graphs (or Social Networks)
Different measures for different types of flow
Used goods (eg books) consider trails in graphs?
Money (eg a dollar note) consider walks in graphs?
Markov process?
Attitude/Belief Influence process
Gossip and infection similarities and differences
Package delivery known destination? Shortest route?
Paths vs. Walks vs. Trails
Transfer vs. Duplication
Centrality Measures in Graphs (or Social Networks)
Some references:
Freeman L.C. (1979): Centrality in social networks: Conceptual clarification.
Social Networks 1, 215-239.
Bonacich P. (1991): Simultaneous group and individual centralities. Social
Networks 13, 155-168
Friedkin N. E. (1991): Theoretical foundations for centrality measures. Amer.
Journal of Sociology 96, 1478-1504
Stephenson K., Zelen M. (1989): Rethinking centrality:methods and examples.
Social Networks 11, 1-37.
Borgatti S.P. (2005): Centrality and network flow. Social Networks 27, 55-71.
Ruhnau B. (2000): Eigenvector centrality a node centrality? Social Networks
22, 357-365
Faust K. (1997): Centrality in affiliation networks. Social Networks 19, 157-191.
Newman M. (2005): A measure of betweeness centrality based on random walks.
Social Networks 27, 39-54.
Bell D., Atkinson J., Carlson J. (1999): Centrality measures for disease
transmission networks. Social Networks 21, 1-21.

Potrebbero piacerti anche