Tetrises and Graph Coloring (joke included) Aneta tastn, Ondej - - PowerPoint PPT Presentation

tetrises and graph coloring
SMART_READER_LITE
LIVE PREVIEW

Tetrises and Graph Coloring (joke included) Aneta tastn, Ondej - - PowerPoint PPT Presentation

Tetrises and Graph Coloring (joke included) Aneta tastn, Ondej plchal ErdsFaberLovsz conjecture - clique version If n complete graphs, each having exactly n vertices, have the property that every pair of complete


slide-1
SLIDE 1

Tetrises and Graph Coloring

(joke included)

Aneta Štastná, Ondřej Šplíchal

slide-2
SLIDE 2

Erdős–Faber–Lovász conjecture

  • clique version
  • If n complete

graphs, each having exactly n vertices, have the property that every pair of complete graphs has at most one shared vertex, then the union of the graphs can be colored with n colors.

  • Example: n = 5
slide-3
SLIDE 3

Tetrises

New Perspective on EFL

slide-4
SLIDE 4

Difgerent perspectives on EFL

slide-5
SLIDE 5

Tetris with maximal number of crossings T

  • Each two cliques intersect in
  • ne vertex.
  • No vertex belongs to more than

2 cliques.

slide-6
SLIDE 6

Coloring of T

slide-7
SLIDE 7

Coloring of T

slide-8
SLIDE 8

Coloring of T

slide-9
SLIDE 9

Coloring of T

slide-10
SLIDE 10

Coloring of T

slide-11
SLIDE 11

Coloring of T

slide-12
SLIDE 12

Coloring of T

slide-13
SLIDE 13

Coloring of T

slide-14
SLIDE 14

Coloring of T

slide-15
SLIDE 15

Coloring of T

slide-16
SLIDE 16

Coloring of T

slide-17
SLIDE 17

Coloring of T

slide-18
SLIDE 18

Coloring of T

slide-19
SLIDE 19

Coloring of T

slide-20
SLIDE 20

Coloring of T

slide-21
SLIDE 21

Coloring of T

slide-22
SLIDE 22

Coloring of T

slide-23
SLIDE 23

Coloring of T

slide-24
SLIDE 24

Coloring of T

slide-25
SLIDE 25

Coloring of T

slide-26
SLIDE 26

Coloring of T

slide-27
SLIDE 27

Coloring of T

slide-28
SLIDE 28

Coloring of T

slide-29
SLIDE 29

Coloring of T

slide-30
SLIDE 30

Coloring of T

slide-31
SLIDE 31

Coloring of T

slide-32
SLIDE 32

EFL and b-coloring of tight bipartite graphs

  • [Lin, Chang (2013)]:

EFL implies that class Bn of tight bipartite graphs is n- or (n-1)-b-colorable.

  • In proof EFL is used as following:

Let G be tight bipartite graph and G* it’s conversion to graph satisfying hypothesis of clique version of EFL. If G* is n colorable, then G is n- or (n-1)-b-colorable.

slide-33
SLIDE 33

EFL and b-coloring of tight bipartite graphs

  • S(Kn) – graph created from Kn by subdivision of all

edges

  • Tetris for S(Kn)* is maximal tetris.
slide-34
SLIDE 34

EFL and b-coloring of tight bipartite graphs

  • n-b-colorability in tetrises: all bricks with one fjlled tile

have distinct colors

  • [Lin, Chang (2013)]:

– S(Kn) is n-b-colorable for n odd – S(Kn) is (n-1)-b-colorable for n even

  • Proof in tetrises:

– n odd: from coloring of maximal tetris – n even: S(Kn) not n-b-colorable and EFL holds for

S(Kn) from tetris coloring algorithm

slide-35
SLIDE 35

EFL and b-coloring of tight bipartite graphs

  • Gn,k – taking S(Kn), merging vertices {1,2}, … , {1,k}

into one, adding some vertices of degree 1

  • In tetrises: Gn,k* is tetris with one brick with k fjlled

tiles, rest are all bricks with two fjlled tiles which can be added (and some bricks with one fjlled tile).

slide-36
SLIDE 36

EFL and b-coloring of tight bipartite graphs

  • [Lin, Chang (2013)]: All graphs in Gn,k are n- or (n-1)-

b-colorable.

  • Proof using tetrises:

– Color tetris of Gn,k* with n colors (alteration of

algorithm for tetris with maximal number of crossings)

– Implies n- or (n-1)-b-colorability by theorem of

[Lin, Chang (2013)]

slide-37
SLIDE 37

EFL and dense graphs

  • deg(x) – number of cliques in which vertex/brick x

belongs

  • Two bricks/vertices collide, if they belong to same

clique

  • [Sánchez-Arroyo (2008)]:

Let k be number of bricks with degree at least deg(x) coliding with x, then

slide-38
SLIDE 38

EFL and dense graphs

  • New result:

– Choose any p bricks. – Denote z number of cliques containing at least one

  • f these chosen bricks.

– Let x be any brick with deg(x) > p+1 – Let k be number of bricks with degree at least

deg(x) coliding with x

– Then – Note that for p = 0 we get result from [Sánchez-

Arroyo (2008)]

slide-39
SLIDE 39

EFL and dense graphs

  • Let d be minimal degree of brick in tetris T and D

maximal degree of a brick in tetris T.

  • Corollary:

– If d > p+1 and – Then T can be colored by at most n colors.

  • Corollary (weaker):

– If d is at least 3 and d times D is at least 2n, then

T can be colored by at most n colors.

slide-40
SLIDE 40

Conclusion

  • We have seen and understood many difgerent

approaches to EFL and it’s connection to other mathematical structures.

  • We have proven results about b-coloring of tight

bipartite graphs in a difgerent and easier way.

  • We have generalized and improved results for dense

graphs.

slide-41
SLIDE 41

References

  • Wu-Hsiung Lin and Gerard J. Chang. b-coloring of

tight bipartite graphs and the Erdős–Faber–Lovász

  • conjecture. Discrete Applied Mathematics,

161(7):1060 – 1066, 2013. ISSN 0166-218X.

  • Abdón Sánchez-Arroyo. The Erdős–Faber–Lovász

conjecture for dense hypergraphs. Discrete Mathematics, 308(5): 991 – 992, 2008. ISSN 0012-

  • 365X. Selected Papers from 20th British Combinatorial

Conference.

slide-42
SLIDE 42

Joke

Thanks for watching!