Locally identifying colorings of graphs Aline Parreau Joint work - - PowerPoint PPT Presentation

locally identifying colorings of graphs
SMART_READER_LITE
LIVE PREVIEW

Locally identifying colorings of graphs Aline Parreau Joint work - - PowerPoint PPT Presentation

Locally identifying colorings of graphs Aline Parreau Joint work with: Louis Esperet, Sylvain Gravier, Micka el Montassier, Pascal Ochem and: Florent Foucaud, Iiro Honkala, Tero Laihonen, Guillem Perarnau Bordeaux Workshop on Identifying


slide-1
SLIDE 1

Locally identifying colorings of graphs

Aline Parreau

Joint work with: Louis Esperet, Sylvain Gravier, Micka¨ el Montassier, Pascal Ochem and: Florent Foucaud, Iiro Honkala, Tero Laihonen, Guillem Perarnau

Bordeaux Workshop on Identifying Codes November 21-25, 2011

1/13

slide-2
SLIDE 2

Identification with colors ?

Identifying coloring of a graph G:

  • c : V → N
  • c(N[x]) = c(N[y]) for any vertices x = y
  • χid(G): minimum number of colors needed to identify G

2 2 3 2 1 1

2/13

slide-3
SLIDE 3

Identification with colors ?

Identifying coloring of a graph G:

  • c : V → N
  • c(N[x]) = c(N[y]) for any vertices x = y
  • χid(G): minimum number of colors needed to identify G

2 2 3 2 1 1

{1, 2, 3} {2, 3} {1, 2, 3} {1, 2} {1, 2, 3} {1, 2}

2/13

slide-4
SLIDE 4

Identification with colors ?

Identifying coloring of a graph G:

  • c : V → N
  • c(N[x]) = c(N[y]) for any vertices x = y
  • χid(G): minimum number of colors needed to identify G

2 2 3 2 1 4

{1, 2, 3} {2, 3} {2, 3, 4} {1, 2} {1, 2, 3, 4} {2, 4}

2/13

slide-5
SLIDE 5

Identification with colors ?

Identifying coloring of a graph G:

  • c : V → N
  • c(N[x]) = c(N[y]) for any vertices x = y
  • χid(G): minimum number of colors needed to identify G

Few remarks:

  • only exists for twin-free graphs (like id-codes)
  • χid(G) ≤ γID(G) + 1

4 4 4 2 1 3

2/13

slide-6
SLIDE 6

Global to local colorings

Identifying coloring of a graph G = (V , E):

  • c : V → N;
  • For any x = y in V , c(N[x]) = c(N[y]);
  • χid(G): minimum number of colors needed to identify G;

Locally identifying coloring (lid-coloring) of a graph G = (V , E):

  • c : V → N, c(x) = c(y) for xy ∈ E;
  • For any xy ∈ E, c(N[x]) = c(N[y]), if possible;
  • χlid(G): min. number of colors needed to locally identify G.

3/13

slide-7
SLIDE 7

Global to local colorings

Identifying coloring of a graph G = (V , E):

  • c : V → N;
  • For any x = y in V , c(N[x]) = c(N[y]);
  • χid(G): minimum number of colors needed to identify G;

Locally identifying coloring (lid-coloring) of a graph G = (V , E):

  • c : V → N, c(x) = c(y) for xy ∈ E;
  • For any xy ∈ E, c(N[x]) = c(N[y]), if possible;
  • χlid(G): min. number of colors needed to locally identify G.

Why?

  • Always exists.
  • Refinment of classic colorings: χ(G) ≤ χlid(G)

3/13

slide-8
SLIDE 8

An example

Def: ∀xy ∈ E, c(x) = c(y) and c(N[x]) = c(N[y])

4/13

slide-9
SLIDE 9

An example

Def: ∀xy ∈ E, c(x) = c(y) and c(N[x]) = c(N[y])

4/13

slide-10
SLIDE 10

An example

Def: ∀xy ∈ E, c(x) = c(y) and c(N[x]) = c(N[y]) 1 2 3 4 2 4 4 2 3 1 4 3 2 3 1 1

χlid(G) ≤ 4

4/13

slide-11
SLIDE 11

An example

Def: ∀xy ∈ E, c(x) = c(y) and c(N[x]) = c(N[y]) 1 1

χlid(G) ? = 4

4/13

slide-12
SLIDE 12

An example

Def: ∀xy ∈ E, c(x) = c(y) and c(N[x]) = c(N[y]) 1 1 2 3

χlid(G) ? = 4

4/13

slide-13
SLIDE 13

An example

Def: ∀xy ∈ E, c(x) = c(y) and c(N[x]) = c(N[y]) 1 1 2 3

χlid(G) = 4 but χ(G) = 3

4/13

slide-14
SLIDE 14

An example

Def: ∀xy ∈ E, c(x) = c(y) and c(N[x]) = c(N[y]) 1 1 2 3

χlid(G) = 4 but χ(G) = 3

For each k, there exists graph Gk s.t χ(Gk) = 3 and χlid(Gk) = k No upper bound with χ !

4/13

slide-15
SLIDE 15

Upper bound on a graph with n vertices ?

Classic colorings: χ(G) = n ⇔ G = Kn Lid-colorings: for which graphs χlid(G) = n?

5/13

slide-16
SLIDE 16

Upper bound on a graph with n vertices ?

Classic colorings: χ(G) = n ⇔ G = Kn Lid-colorings: for which graphs χlid(G) = n?

  • Kn
  • Pk−1

2k

: 1 2 3 4 5 6

Extremal graph for identifying codes !

5/13

slide-17
SLIDE 17

Upper bound on a graph with n vertices ?

Classic colorings: χ(G) = n ⇔ G = Kn Lid-colorings: for which graphs χlid(G) = n?

  • Kn
  • Pk−1

2k

: 1 2 3 4 5 6

Extremal graph for identifying codes !

  • ... ?

Caracterize graphs G such that χlid(G) = n. Open question

5/13

slide-18
SLIDE 18

Maximum degree

Classic colorings: χ(G) ≤ ∆ + 1, tight Lid-colorings:

  • χlid(G) ≤ χ(G 3) ≤ ∆3 − ∆2 + ∆ + 1
  • Graphs with χlid(G) ≥ ∆2 − ∆ + 1

6/13

slide-19
SLIDE 19

Maximum degree

Classic colorings: χ(G) ≤ ∆ + 1, tight Lid-colorings:

  • χlid(G) ≤ χ(G 3) ≤ ∆3 − ∆2 + ∆ + 1
  • Graphs with χlid(G) ≥ ∆2 − ∆ + 1

6/13

slide-20
SLIDE 20

Maximum degree

Classic colorings: χ(G) ≤ ∆ + 1, tight Lid-colorings:

  • χlid(G) ≤ χ(G 3) ≤ ∆3 − ∆2 + ∆ + 1
  • Graphs with χlid(G) ≥ ∆2 − ∆ + 1

For any graph G with ∆ ≥ 3: χlid(G) ≤ 2∆2 − 3∆ + 3 Theorem (Foucaud,Honkala,Laihonen,P.,Perarnau, 2011+) Do we always have χlid(G) ≤ ∆2 + O(∆) ? Open question

6/13

slide-21
SLIDE 21

Bipartite graphs: the paths

With 4 colors :

7/13

slide-22
SLIDE 22

Bipartite graphs: the paths

With 4 colors : 1 2 3 4 1 2 3 4

7/13

slide-23
SLIDE 23

Bipartite graphs: the paths

With 4 colors : 1 2 3 4 1 2 3 4

1, 2 1, 2, 3 2, 3, 4 1, 3, 4 1, 2, 4 1, 2, 3 2, 3, 4 3, 4

So: χlid(Pk) ≤ 4

7/13

slide-24
SLIDE 24

Bipartite graphs: the paths

With 4 colors : 1 2 3 4 1 2 3 4

1, 2 1, 2, 3 2, 3, 4 1, 3, 4 1, 2, 4 1, 2, 3 2, 3, 4 3, 4

So: χlid(Pk) ≤ 4 Is it possible with 3 colors ?

7/13

slide-25
SLIDE 25

Bipartite graphs: the paths

With 4 colors : 1 2 3 4 1 2 3 4

1, 2 1, 2, 3 2, 3, 4 1, 3, 4 1, 2, 4 1, 2, 3 2, 3, 4 3, 4

So: χlid(Pk) ≤ 4 Is it possible with 3 colors ? 1 2

1, 2

7/13

slide-26
SLIDE 26

Bipartite graphs: the paths

With 4 colors : 1 2 3 4 1 2 3 4

1, 2 1, 2, 3 2, 3, 4 1, 3, 4 1, 2, 4 1, 2, 3 2, 3, 4 3, 4

So: χlid(Pk) ≤ 4 Is it possible with 3 colors ? 1 2

1, 2

3

1, 2, 3

7/13

slide-27
SLIDE 27

Bipartite graphs: the paths

With 4 colors : 1 2 3 4 1 2 3 4

1, 2 1, 2, 3 2, 3, 4 1, 3, 4 1, 2, 4 1, 2, 3 2, 3, 4 3, 4

So: χlid(Pk) ≤ 4 Is it possible with 3 colors ? 1 2

1, 2

3

1, 2, 3

2

2, 3

7/13

slide-28
SLIDE 28

Bipartite graphs: the paths

With 4 colors : 1 2 3 4 1 2 3 4

1, 2 1, 2, 3 2, 3, 4 1, 3, 4 1, 2, 4 1, 2, 3 2, 3, 4 3, 4

So: χlid(Pk) ≤ 4 Is it possible with 3 colors ? 1 2

1, 2

3

1, 2, 3

2

2, 3

1

1, 2, 3

2 3 2

1, 2 1, 2, 3 2, 3 2, 3

χlid(Pk) = 3 ⇔ k is odd ... χlid is not heriditary !

7/13

slide-29
SLIDE 29

Bipartite graphs

L0 L1 L2 L3 L4

8/13

slide-30
SLIDE 30

Bipartite graphs

L0 L1 L2 L3 L4 1 2 3 4 1 → → → → →

1, 2 1, 2, 3 2, 3, 4 or 2, 3 1, 3, 4 or 3, 4 1, 4

  • 3 ≤ χlid(B) ≤ 4
  • To decide between 3 and 4 is NP-complete (reduction from

2-coloring of hypergraph)

  • Polynomial for trees, grids and hypercubes (χlid = 3), regular

bipartite graphs...

8/13

slide-31
SLIDE 31

Perfect Graphs Perfect

Chordal Permutation Line of bipartite Cograph Trees k-trees Split Bipartite Interval

9/13

slide-32
SLIDE 32

Perfect Graphs Perfect

Chordal Permutation Line of bipartite Cograph Trees k-trees Split Bipartite Interval Trees Bipartite ≤ 4 = 2χ ≤ 4 = 2χ

9/13

slide-33
SLIDE 33

Perfect Graphs Perfect

Chordal Permutation Line of bipartite Cograph Trees k-trees Split Bipartite Interval Trees Bipartite ≤ 4 = 2χ ≤ 4 = 2χ k-trees

≤ 2χ

Split Interval Cograph

≤ 2χ ≤ 2χ ≤ 2χ

Perfect

Chordal Not bounded by χ ? Do we have χlid(G) ≤ 2χ(G) for a chordal graph G? Open question

9/13

slide-34
SLIDE 34

Planar Graphs

Planar graphs:

  • Worse example : 8 colors,
  • Really large (1000 ?) bound by

Gonzcales and Pinlou P3

8

10/13

slide-35
SLIDE 35

Planar Graphs

Planar graphs:

  • Worse example : 8 colors,
  • Really large (1000 ?) bound by

Gonzcales and Pinlou

  • With large girth (36) bounded by 5

P3

8

10/13

slide-36
SLIDE 36

Planar Graphs

Planar graphs:

  • Worse example : 8 colors,
  • Really large (1000 ?) bound by

Gonzcales and Pinlou

  • With large girth (36) bounded by 5

P3

8

10/13

slide-37
SLIDE 37

Planar Graphs

Planar graphs:

  • Worse example : 8 colors,
  • Really large (1000 ?) bound by

Gonzcales and Pinlou

  • With large girth (36) bounded by 5

P3

8

Outerplanar graphs:

  • General bound: 20 colors,
  • Max outerplanar graphs: ≤ 6 colors,
  • Without triangles: ≤ 8 colors,
  • Examples with at most 6 colors

P2

6

Do we have χlid(G) ≤ 8 for planar graphs and χlid(G) ≤ 6 for outer- planar graphs ? Open question

10/13

slide-38
SLIDE 38

A remark

  • For some subclasses of perfect graphs : χlid(G) ≤ 2ω(G) = 2χ(G)
  • For planar graphs, worse example : χlid(G) ≤ 8 = 2χ(G)
  • For outerplanar graphs, worse example : χlid(G) ≤ 6 = 2χ(G)
  • ...

For which graphs do we have χlid(G) ≤ 2χ(G) ? Open question

11/13

slide-39
SLIDE 39

Another remark

  • χlid(G) = 2 ⇔ G = K2
  • χlid(G) = 3 ⇒ G = K3 or G is bipartite
  • χlid(G) = 3 and χ(G) = 3 ⇔ G = K3

Caracterize graphs G such that χlid(G) = χ(G). Are they only the complete graphs ? Open question

12/13

slide-40
SLIDE 40

Conclusion

Lot of open questions:

  • Graphs with χlid = n ?
  • Graphs with χlid = χ ?
  • Do we have χlid(G) ≤ ∆2 + O(∆) ?
  • Do we have χlid(G) ≤ 2χ(G) for chordal graphs? for planar graphs?

for which graphs?

  • Find a good bound for planar graphs.
  • Find a ”nice” application of lid-colorings

Thanks !

13/13