Vous êtes sur la page 1sur 3

Applied Algorithms : Chordal Graphs

Hemanshu Vadehra
Roll No. : 10305060
September 7, 2010

Abstract
A few problems on chordal graphs and it’s subclasses, including a special section
on the real-world applications of such graphs.

1 Chordless cycles in non-chordal graphs


To find a chordless cycle of size >= 4 in linear time in a non-chordal graph, we can
do the following:
1. Perform Lex-BFS to get an ordering of vertices and take its reverse (this
would have been a PEO if the graph was chordal).
2. For each vi starting with vn , and proceeding in descending order:
a. If vi has neighbours vj and vk , such that there is no edge from vj to
vk , then vi , vj and vk are part of a chordless cycle of size >= 4.
b. Find a path from vj to vk not including vi , which is a set of vertices
say, S.
c. The set of vertices S ∪ {vi , vj , vk } is a chordless cycle of size >= 4.
It can be shown that this algorithm will run in O(n + m) time since we know Lex-
BFS completes in linear time and in step 2 we onnly need to consider neighbours
vj and vk of vi such that j, k <= i.

2 Recognizing split graphs


In a split graph G, if we partition the vertices into sets C (a maximal clique) and
I (an independent set), then all vertices in C have degree >= |C| − 1 (since they
are adjacent to every other vertex in C). Also all vertices in I are only adjacent to
vertices in C. So the sum of the degrees of all vertices in C is equal to |C| × (|C| − 1)
and the sum of degrees of every vertex in I. Therefore split graphs can be recognized
from their degree sequences. Let the degree sequence of G be d1 >= d2 >= ... >=
dk , and let m be the largest value of i such that di >= i − 1 (thus m is the number
of vertices in the maximal clique). Then G is a split graph if
m
X k
X
di = m(m − 1) + dj
i=1 j=m+1

1
If this is the case, then the m vertices with the largest degrees form a maximum
clique in G, and the remaining vertices constitute an independent set.

3 Max independent set in weighted chordal


graphs
For weighted versions of these problems, the representation as the intersection graph
of subtrees of a tree is needed. However, this is also implicitly given by the perfect
elimination ordering, where the parent of vertex i is the smallest vertex in N + [i].
The subtree corresponding to i contains i and all vertices in N − [i]. The underlying
tree is assumed to be rooted at vertex n − 1. The descendants of i in the tree are
all those vertices that are reachable from i by a monotonically decreasing path. Let
G[i] denote the subgraph of G induced by descendants of i and let S1 [i] be the
weight of maximum weight independent set in G[i] that excludes i and S2 [i] the
weight of maximum weight independent set including i. Let S[i] be the maximum
of S1 [i], S2 [i]. A child of i is a vertex j < i that is adjacent to i but not to any
vertex k, j < k < i. Let C[i] denote the set of children of i.
Then we have the recurrences[1]:

X
S1 [i] = S[j]
j∈C[i]
X
S2 [i] = wt(i) + S1 [i] + (S1 [j] − S[j])
j∈N − [i]

(
S1 [i] if S1 [i] >= S2 [i]
S[i] =
S2 [i] if S1 [i] < S2 [i].

4 Interval ⊂ Path ⊂ Chordal


An asteroidal triple is a set of three vertices such that each pair is connected by a
path that does not include the neighbours of the third vertex. A path graph that
contains an asteroidal triple is not an interval graph.

A path graph that is not an interval graph[2].

A strong asteroidal triple is an asteroidal triple such that each pair is connected
by a strong path (a path between two non-adjacent vertices that have a common
neighbour). A graph that is chordal and contains a strong asteroidal triple is not a
path graph.

2
A chordal graph that is not a path graph.

5 Applications in real life


One of the most interesting applications of abstract graph theory and combinatorics
would be to select the Indian national football team, since no combination of play-
ers that the federation comes up with ever qualifies for the World Cup. Perhaps
mathematics and computers have the answer!

So we use the following formula :

n n!
r = n Cr = r! (n−r)!

and get :

Team sheet
Goalkeeper GK Subrata Pal
LB Mahesh Gawli
DC Deepak Kumar Mondal
Defenders
DC Irungbam Surkumar Singh
RB Samir Subash Naik
MC P Renedy Singh
Midfielders MC Climax Lawrence
MC Naduparambil Pappachan Pradeep
ST Steven Dias
Strikers ST Baichung Bhutia
ST Sunil Chhetri

References
[1] AA Diwan. Algorithmic problems on chordal graphs. August 2010.
[2] Rafael Villarroel. Combinatorial graphs in tikz.

Vous aimerez peut-être aussi