2ème Cours Cours MPRI 2010 2011habib/Documents/Cours_2_2010.pdfthe memory word GRAF[x,y] Therefore...

24
2` eme Cours Cours MPRI 2010–2011 2` eme Cours Cours MPRI 2010–2011 Michel Habib [email protected] http://www.liafa.jussieu.fr/ ~ habib Chevaleret, septembre 2010

Transcript of 2ème Cours Cours MPRI 2010 2011habib/Documents/Cours_2_2010.pdfthe memory word GRAF[x,y] Therefore...

Page 1: 2ème Cours Cours MPRI 2010 2011habib/Documents/Cours_2_2010.pdfthe memory word GRAF[x,y] Therefore the query : xy ∈E? Can be done in 2 tests O(1). 2`eme Cours Cours MPRI 2010–2011

2eme Cours Cours MPRI 2010–2011

2eme CoursCours MPRI 2010–2011

Michel [email protected]

http://www.liafa.jussieu.fr/~habib

Chevaleret, septembre 2010

Page 2: 2ème Cours Cours MPRI 2010 2011habib/Documents/Cours_2_2010.pdfthe memory word GRAF[x,y] Therefore the query : xy ∈E? Can be done in 2 tests O(1). 2`eme Cours Cours MPRI 2010–2011

2eme Cours Cours MPRI 2010–2011

Schedule

Comments on last course

Chordal graphsLexicographic Breadth First Search LexBFSSimplicial elimination scheme

Exercices

Page 3: 2ème Cours Cours MPRI 2010 2011habib/Documents/Cours_2_2010.pdfthe memory word GRAF[x,y] Therefore the query : xy ∈E? Can be done in 2 tests O(1). 2`eme Cours Cours MPRI 2010–2011

2eme Cours Cours MPRI 2010–2011

Comments on last course

Fagin’s Theorems again

Fagin’s theorems in structural complexity

Characterizations without any notion of machines or algorithms !

NP

The class of all graph-theoretic properties expressible in existentialsecond-order logic is precisely NP.

P

The class of all graph-theoretic properties expressible in Hornexistential second-order logic with successor is precisely P.

Page 4: 2ème Cours Cours MPRI 2010 2011habib/Documents/Cours_2_2010.pdfthe memory word GRAF[x,y] Therefore the query : xy ∈E? Can be done in 2 tests O(1). 2`eme Cours Cours MPRI 2010–2011

2eme Cours Cours MPRI 2010–2011

Comments on last course

Quadratic space in linear time

◮ Select a 2-dimensional array GRAF of size n2

construct an auxillary unidimensional array of size m EDGE :For j=1 to mxy being the j th edge of GGRAF [x , y ] = jEDGE [j] = a pointer to the memory word GRAF [x , y ]

◮ The construction of the EDGE array requires O(m) time

◮ Memory used n2 +m ∈ O(n2)

Page 5: 2ème Cours Cours MPRI 2010 2011habib/Documents/Cours_2_2010.pdfthe memory word GRAF[x,y] Therefore the query : xy ∈E? Can be done in 2 tests O(1). 2`eme Cours Cours MPRI 2010–2011

2eme Cours Cours MPRI 2010–2011

Comments on last course

◮ xy ∈ E iff EDGE [GRAF [x , y ]] contains a pointer pointing tothe memory word GRAF [x , y ]

◮ Therefore the query : xy ∈ E ?Can be done in 2 tests O(1).

Page 6: 2ème Cours Cours MPRI 2010 2011habib/Documents/Cours_2_2010.pdfthe memory word GRAF[x,y] Therefore the query : xy ∈E? Can be done in 2 tests O(1). 2`eme Cours Cours MPRI 2010–2011

2eme Cours Cours MPRI 2010–2011

Comments on last course

Any graph solution for the rectangle problem ?

Page 7: 2ème Cours Cours MPRI 2010 2011habib/Documents/Cours_2_2010.pdfthe memory word GRAF[x,y] Therefore the query : xy ∈E? Can be done in 2 tests O(1). 2`eme Cours Cours MPRI 2010–2011

2eme Cours Cours MPRI 2010–2011

Chordal graphs

A nice graph

Start with the graph of the planar tiling and keep exactly 2integers edges by rectangle.This yields a graph (possibly with parallel edges) in whichall vertices have even degrees except the corners.Take a maximal path starting in one corner ....

Page 8: 2ème Cours Cours MPRI 2010 2011habib/Documents/Cours_2_2010.pdfthe memory word GRAF[x,y] Therefore the query : xy ∈E? Can be done in 2 tests O(1). 2`eme Cours Cours MPRI 2010–2011

2eme Cours Cours MPRI 2010–2011

Chordal graphs

Lexicographic Breadth First Search LexBFS

Lexicographic Breadth First Search (LexBFS)

Data: a graph G = (V ,E ) and a start vertex s

Result: an ordering σ of V

Assign the label ∅ to all verticeslabel(s)← {n}for i ← n a 1 do

Pick an unumbered vertex v with lexicographically largest labelσ(i)← vforeach unnumbered vertex w adjacent to v do

label(w)← label(w).{i}end

end

Page 9: 2ème Cours Cours MPRI 2010 2011habib/Documents/Cours_2_2010.pdfthe memory word GRAF[x,y] Therefore the query : xy ∈E? Can be done in 2 tests O(1). 2`eme Cours Cours MPRI 2010–2011

2eme Cours Cours MPRI 2010–2011

Chordal graphs

Lexicographic Breadth First Search LexBFS

1

76

5

4

3

2The ordering of the LexBFS search is 7,6,5,4,3,2,1. Note that thereverse ordering is not simplicial, since G is not chordal

Page 10: 2ème Cours Cours MPRI 2010 2011habib/Documents/Cours_2_2010.pdfthe memory word GRAF[x,y] Therefore the query : xy ∈E? Can be done in 2 tests O(1). 2`eme Cours Cours MPRI 2010–2011

2eme Cours Cours MPRI 2010–2011

Chordal graphs

Lexicographic Breadth First Search LexBFS

It is just a breadth first search with a tie break rule.We are now considering a characterization of the

order in which a LexBFS explores the vertices.

Page 11: 2ème Cours Cours MPRI 2010 2011habib/Documents/Cours_2_2010.pdfthe memory word GRAF[x,y] Therefore the query : xy ∈E? Can be done in 2 tests O(1). 2`eme Cours Cours MPRI 2010–2011

2eme Cours Cours MPRI 2010–2011

Chordal graphs

Lexicographic Breadth First Search LexBFS

Property (LexB)

an order σ on V , if a < b < c and ac ∈ E but ab /∈ E , then itexists a vertex d such that d < a and db ∈ E and dc /∈ E .

d cba

Theorem

For a graph G = (V ,E ), an order σ on V is a LexBFS of G iff σsatisfies property (LexB).

Page 12: 2ème Cours Cours MPRI 2010 2011habib/Documents/Cours_2_2010.pdfthe memory word GRAF[x,y] Therefore the query : xy ∈E? Can be done in 2 tests O(1). 2`eme Cours Cours MPRI 2010–2011

2eme Cours Cours MPRI 2010–2011

Chordal graphs

Lexicographic Breadth First Search LexBFS

4 points condition

Questions◮ Under which condition an order σ on V correspond to some

graph search ?

◮ What are the properties of these orderings ?

Main reference :

D.G. Corneil et R. M. Krueger, A unified view of graph searching,SIAM J. Discrete Math, 22, N 4 (2008) 1259-1276

Page 13: 2ème Cours Cours MPRI 2010 2011habib/Documents/Cours_2_2010.pdfthe memory word GRAF[x,y] Therefore the query : xy ∈E? Can be done in 2 tests O(1). 2`eme Cours Cours MPRI 2010–2011

2eme Cours Cours MPRI 2010–2011

Chordal graphs

Lexicographic Breadth First Search LexBFS

A characterisation theorem for chordal graphs

Theorem

Dirac 1961, Fulkerson, Gross 1965, Gavril 1974, Rose, Tarjan,Lueker 1976.

(0) G is chordal (every cycle of length ≥ 4 has a chord) .

(i) G has a simplicial elimination scheme

(ii) Every minimal separator is a clique

Page 14: 2ème Cours Cours MPRI 2010 2011habib/Documents/Cours_2_2010.pdfthe memory word GRAF[x,y] Therefore the query : xy ∈E? Can be done in 2 tests O(1). 2`eme Cours Cours MPRI 2010–2011

2eme Cours Cours MPRI 2010–2011

Chordal graphs

Simplicial elimination scheme

Simplicial

5

1 4 38

6 7 2

A vertex is simplicial if its neighbourhood is a clique.

Simplicial elimination scheme

σ = [x1 . . . xi . . . xn] is a simplicial elimination scheme if xi issimplicial in the subgraph Gi = G [{xi . . . xn}]

ca b

Page 15: 2ème Cours Cours MPRI 2010 2011habib/Documents/Cours_2_2010.pdfthe memory word GRAF[x,y] Therefore the query : xy ∈E? Can be done in 2 tests O(1). 2`eme Cours Cours MPRI 2010–2011

2eme Cours Cours MPRI 2010–2011

Chordal graphs

Simplicial elimination scheme

Minimal Separators

A subset of vertices S is a minimal separator if Gif there exist a, b ∈ G such that a and b are not connected inG − S .and S is minimal for inclusion with this property .

Page 16: 2ème Cours Cours MPRI 2010 2011habib/Documents/Cours_2_2010.pdfthe memory word GRAF[x,y] Therefore the query : xy ∈E? Can be done in 2 tests O(1). 2`eme Cours Cours MPRI 2010–2011

2eme Cours Cours MPRI 2010–2011

Chordal graphs

Simplicial elimination scheme

Chordal graphs are hereditary

Page 17: 2ème Cours Cours MPRI 2010 2011habib/Documents/Cours_2_2010.pdfthe memory word GRAF[x,y] Therefore the query : xy ∈E? Can be done in 2 tests O(1). 2`eme Cours Cours MPRI 2010–2011

2eme Cours Cours MPRI 2010–2011

Chordal graphs

Simplicial elimination scheme

Theorem [Tarjan et Yannakakis, 1984]

G is a chordal graph iff every LexBFS ordering provides a simplicialelimination scheme.

1

1 8

7

6

5

4

32

Page 18: 2ème Cours Cours MPRI 2010 2011habib/Documents/Cours_2_2010.pdfthe memory word GRAF[x,y] Therefore the query : xy ∈E? Can be done in 2 tests O(1). 2`eme Cours Cours MPRI 2010–2011

2eme Cours Cours MPRI 2010–2011

Chordal graphs

Simplicial elimination scheme

How can we prove such an algorithmic theorem ?

1. A direct proof, finding the invariants ?

2. Find some structure of chordal graphs

3. Understand how LexBFS explores a chordal graph

Page 19: 2ème Cours Cours MPRI 2010 2011habib/Documents/Cours_2_2010.pdfthe memory word GRAF[x,y] Therefore the query : xy ∈E? Can be done in 2 tests O(1). 2`eme Cours Cours MPRI 2010–2011

2eme Cours Cours MPRI 2010–2011

Chordal graphs

Simplicial elimination scheme

A direct proof

Theorem [Tarjan et Yannakakis, 1984]

G is a chordal graph iff every LexBFS ordering provides a simplicialelimination scheme.

Page 20: 2ème Cours Cours MPRI 2010 2011habib/Documents/Cours_2_2010.pdfthe memory word GRAF[x,y] Therefore the query : xy ∈E? Can be done in 2 tests O(1). 2`eme Cours Cours MPRI 2010–2011

2eme Cours Cours MPRI 2010–2011

Chordal graphs

Simplicial elimination scheme

Demonstration.

Let c be the leftmost non simplicial vertex.Therefore it exists a < b ∈ N(c) with ab /∈ E . Using LexBproperty, it necessarily exists d < a with db ∈ E and dc /∈ E .Since G is chordal, we have ad /∈ E (else we would have the cycle[a, c , b, d ] without a chord).But then considering the triple d , a, b, it exists d ′ < d such thatd ′a ∈ E and d ′b /∈ E .If dd ′ ∈ E , using the cycle [d , d ′, a, c , b] we must have the chordd ′c ∈ E which provides the cycle [d , d ′c , b] which has no chord.Therefore dd ′ /∈ E .And we construct an infinite sequence of such d and d’.

Page 21: 2ème Cours Cours MPRI 2010 2011habib/Documents/Cours_2_2010.pdfthe memory word GRAF[x,y] Therefore the query : xy ∈E? Can be done in 2 tests O(1). 2`eme Cours Cours MPRI 2010–2011

2eme Cours Cours MPRI 2010–2011

Chordal graphs

Simplicial elimination scheme

Consequences

G has a linear number of maximal cliques.Computing a maximum clique ω(G ) is polynomial.Computing χ(G ) also

Page 22: 2ème Cours Cours MPRI 2010 2011habib/Documents/Cours_2_2010.pdfthe memory word GRAF[x,y] Therefore the query : xy ∈E? Can be done in 2 tests O(1). 2`eme Cours Cours MPRI 2010–2011

2eme Cours Cours MPRI 2010–2011

Chordal graphs

Simplicial elimination scheme

Let C(x , y) be the set of maximal cliques that contain x and y.

Clique Consecutivity property

In a lexBFS ordering τ if y is the first vertex after x s.t.(C (x , y)) = {C}, then the elements of C visited after x areconsecutive in τ .

Page 23: 2ème Cours Cours MPRI 2010 2011habib/Documents/Cours_2_2010.pdfthe memory word GRAF[x,y] Therefore the query : xy ∈E? Can be done in 2 tests O(1). 2`eme Cours Cours MPRI 2010–2011

2eme Cours Cours MPRI 2010–2011

Exercices

Helly Property

Definition

A subset family {Ti}i∈I satisfies Helly property ifJ ⊆ I et ∀i , j ∈ J Ti ∩ Tj 6= ∅ implies ∩i ∈JTi 6= ∅

Exercise

Subtrees in a tree satisfy Helly property.

Page 24: 2ème Cours Cours MPRI 2010 2011habib/Documents/Cours_2_2010.pdfthe memory word GRAF[x,y] Therefore the query : xy ∈E? Can be done in 2 tests O(1). 2`eme Cours Cours MPRI 2010–2011

2eme Cours Cours MPRI 2010–2011

Exercices

Classes of twin vertices

Definition

x and y are called false twins, (resp. true twins) ifN(x) = N(y) (resp. N(x) ∪ {x} = N(y) ∪ {y}))