Line graphs, triangle graphs, and further generalizations Aparna - - PowerPoint PPT Presentation

line graphs triangle graphs and further generalizations
SMART_READER_LITE
LIVE PREVIEW

Line graphs, triangle graphs, and further generalizations Aparna - - PowerPoint PPT Presentation

Line graphs, triangle graphs, and further generalizations Aparna Lakshmanan S. 1 , Csilla Bujts 2 , Zsolt Tuza 2,3 1 St. Xavier's College for Women, Aluva, India 2 University of Pannonia, Veszprm, Hungary 3 Rnyi Institute, Hungarian Academy


slide-1
SLIDE 1

Line graphs, triangle graphs, and further generalizations

Aparna Lakshmanan S.1, Csilla Bujtás2, Zsolt Tuza2,3

1 St. Xavier's College for Women, Aluva, India 2 University of Pannonia, Veszprém, Hungary 3 Rényi Institute, Hungarian Academy of Sciences, Budapest, Hungary

slide-2
SLIDE 2

Outline

  • 1. Line graphs, -line graphs
  • 2. Characterization: □ is a triangle graph
  • 3. Characterization: □ is a -line graph
  • 4. Anti-Gallai graphs, -anti-Gallai-graphs
  • 5. Recognizing -line graphs and -anti-Gallai graphs: NP-complete for

every ≥ 3

slide-3
SLIDE 3
  • Line graph ()

vertices of () ↔ edges (-subgraphs) of edges of () ↔ two edges share a in

  • Triangle graph ()

vertices of () ↔ triangles (-subgraphs) of edges of () ↔ two triangles share an edge () in

slide-4
SLIDE 4
  • Line graph ()

vertices of () ↔ edges (-subgraphs) of edges of () ↔ two edges share a in

  • Triangle graph ()

vertices of () ↔ triangles (-subgraphs) of edges of () ↔ two triangles share an edge () in

  • line graph ()

vertices of () ↔ -subgraphs of edges of () ↔ two -subgraphs share − vertices ( ) in

slide-5
SLIDE 5
  • Line graph ()

vertices of () ↔ edges (-subgraphs) of edges of () ↔ two edges share a in

  • Triangle graph ()

vertices of () ↔ triangles (-subgraphs) of edges of () ↔ two triangles share an edge () in

  • line graph ()

vertices of () ↔ -subgraphs of edges of () ↔ two -subgraphs share − vertices ( ) in () = |V(G)|, () = (), () = ()

slide-6
SLIDE 6

Triangle graph

slide-7
SLIDE 7

Triangle graph

slide-8
SLIDE 8

Triangle graph

slide-9
SLIDE 9

Triangle graph

slide-10
SLIDE 10

Triangle graph

slide-11
SLIDE 11

Triangle graph

slide-12
SLIDE 12

Triangle graph

slide-13
SLIDE 13

Triangle graph – some observations

= ()

  • is , -free
  • If contains a 4 1234 with a preimage 4 in ’, then each further

vertex of is adjacent to 0 or 2 vertices from 1234

  • If contains a diamond , then there is a further vertex adjacent to exactly

three vertices of = (’) & has no 4-component: ’ is -free is diamond-free

slide-14
SLIDE 14
  • line graphs – analogous observations

= (’)

  • is ,-free
  • If contains a subgraph with a preimage in ’, then each

further vertex of is adjacent to 0 or 2 vertices of

  • If contains a diamond , then there are further − 2 vertices of forming

together with three vertices of = (’) & has no -component (k ≥ 2) : ’ is -free is diamond-free

slide-15
SLIDE 15

Cartesian product

:

  • (□) = () × (),
  • (, ), (′, ′) are adjacent iff

[ = & ′ ∈ ()] or [′ ∈ () & = ′]

slide-16
SLIDE 16

Cartesian product – as a triangle graph

is a triangle graph if and only if = and is the line graph of a -free graph (or vice versa). Proof.

  • 1. If and are non-complete → ∃ induced paths 123 and 123

→ (2, 2) has four independent neighbors → is not a triangle-graph

  • 2. = () → , claw-free and diamond-free → , are line graphs
  • 3. = (’)

if = 3, then let ’ = , if ≠ 3, is diamond-free → ’ is K3-free

  • 4. Sufficiency: = (’) & ’ is -free → (’ V ) = (’) □ = □

( and are connected, non-edgeless graphs)

slide-17
SLIDE 17

Cartesian products and k-line graphs

is a -line graph if and only if ∃ ,

  • is the -line graph of a -free graph,
  • is the -line graph of a -free graph, and
  • + ≤

( and are non-edgeless connected graphs)

slide-18
SLIDE 18

Cartesian products and k-line graphs

is a -line graph if and only if ∃ ,

  • is the -line graph of a -free graph,
  • is the -line graph of a -free graph, and
  • + ≤

( and are non-edgeless connected graphs) Sufficiency:

  • Let = (’), = (’), ’ is -free , ’ is -free

(’ ∨ ’) =

  • If − ( + ) = > 0, then ( ’ ∨ ’ ∨ ) =
slide-19
SLIDE 19

Cartesian products and k-line graphs

Necessity:

  • Lemma1: = (’) and contains an induced 4-cycle:

for the preimages in ’: ∖ = ∖ .

slide-20
SLIDE 20

Cartesian products and k-line graphs

Necessity:

  • Lemma 1: = (’) and contains an induced :

for the preimages in ’: ∖ = ∖ . C1 C2 C4 C3 c1 c4 c2 c3

slide-21
SLIDE 21

Cartesian products and k-line graphs

Necessity: Let = = (’). For a copy of consider the preimage -cliques (, ) of ’, and define the sets of the universal and non-universal vertices:

  • = ⋂{(, ): ∈ ()},

= ⋃ , : ∈

slide-22
SLIDE 22

Cartesian products and k-line graphs

Necessity: Let = = (’). For a copy of consider the preimage -cliques (, ) of ’, and define the sets of the universal and non-universal vertices:

  • = ⋂{(, ): ∈ ()},

= ⋃ , : ∈

  • Lemma 2:
  • The non-universal vertices are the same:

=

  • If ∈() → ∃ ,

= ∖ {} ∪ {}

slide-23
SLIDE 23

Cartesian products and k-line graphs

Necessity: Let = = (’). For a copy of consider the preimage -cliques (, ) of ’, and define the sets of the universal and non-universal vertices:

  • = ⋂{(, ): ∈ ()},

= ⋃ , : ∈

  • Lemma 2:
  • The non-universal vertices are the same:

=

  • If ∈() → ∃ ,

= ∖ {} ∪ {}

To complete the proof: = |

| and = | | →

  • is the ( − )-line graph of a -free graph, and
  • is the ( − )-line graph of a - free graph.
  • + ≥
slide-24
SLIDE 24
  • anti-Gallai graph Δ()

vertices of Δ() ↔ -subgraphs of edges of Δ() ↔ two -subgraphs (share a and) contained in a common in

  • Gallai graph Γ()

vertices of Γ() ↔ -subgraphs of edges of Γ() ↔ two -subgraphs share a and NOT contained in a common in

Subgraphs of k-line graphs

  • Anti-Gallai graph Δ()

vertices of Δ() ↔ edges (-subgraphs) of edges of Δ() ↔ the two edges (share a and) are in a common in G

  • Δ() = , Δ() = Δ (); Γ() = ̅, Γ() = Γ()
slide-25
SLIDE 25

Recognizing triangle graphs

The following problems are NP-complete:

  • Recognizing triangle graphs
  • Deciding whether a given graph is the triangle graph of a -free graph

Anand, Escuardo, Gera, Hartke, Stolee (2012):

  • Deciding whether a given connected graph is the anti-Gallai graph of a
  • free graph --- NP-complete problem
slide-26
SLIDE 26

Recognizing triangle graphs

The following problems are NP-complete:

  • Recognizing triangle graphs
  • Deciding whether a given graph is the triangle graph of a -free graph

Proof: Lemma1: If = Δ(’) and is connected, then ’ -free (1) every maximal clique of is a triangle and any two triangles share at most one vertex Lemma2: If = (’) and is connected, ≠ , then ’ is -free (2) each vertex of is contained in at most three maximal cliques, and these are edge-disjoint

  • Given a connected instance

→ Check (1) → If it holds, construct the clique graph = () → satisfies (2) + conn → = Δ() = () (suppose: is ‘triangle-restricted’)

slide-27
SLIDE 27

Recognizing k-line graphs

The following problems are NP-complete for each ≥ 4

  • Recognizing -line graphs
  • Deciding whether a given graph is the -line graph of a -free graph

Proof:

  • = 3 → NP-complete
  • Induction on :

is the -line graph of a -free graph ↔ is the + 1 -line graph of a -free graph (on the class of connected graphs)

  • For = 1 → trivial
  • For = 2 → polynomial-time (Beineke), linear-time (Lehot; Roussopoulos)
  • For = 3 → we have proved: NP-complete
slide-28
SLIDE 28

Recognizing k-anti-Gallai graphs

The following problems are NP-complete for each ≥ 3

  • Recognizing -anti-Gallai graphs
  • Deciding whether a given graph is the -anti-Gallai graph of a -free

graph

  • For = 1 → trivial
  • For = 2 → NP-complete (Anand et al.)
slide-29
SLIDE 29

Recognizing -anti-Gallai graphs

Induction on Given a connected instance → Check diamond-freeness → If it holds, construct ∗ → = Δ(’) ↔ G∗ = Δ(’ ∨ 2) Proof: Lemma: G is -free ↔ Δ() is diamond-free ↔ each maximal clique of Δ() is either an isolated vertex or a ( + 1)-clique, and any two maximal cliques intersect in at most one vertex. G1 G2

slide-30
SLIDE 30

Recognizing -anti-Gallai graphs

G1 G2 Induction on Given a connected instance → Check diamond-freeness → If it holds, construct ∗ → = Δ(’) ↔ G∗ = Δ(’ ∨ 2) Proof: Lemma: G is -free ↔ Δ() is diamond-free ↔ each maximal clique of Δ() is either an isolated vertex or a ( + 1)-clique, and any two maximal cliques intersect in at most one vertex.

slide-31
SLIDE 31

Recognizing -anti-Gallai graphs

Induction on k Given a connected instance G → Check diamond-freeness → If it holds, construct G* → G = Δk(G’) ↔ G*=Δk+1(G’ V 2K1) G1 G2 Proof: Lemma: G is -free ↔ Δ() is diamond-free ↔ each maximal clique of Δ() is either an isolated vertex or a ( + 1)-clique, and any two maximal cliques intersect in at most one vertex.

slide-32
SLIDE 32

Recognizing -anti-Gallai graphs

Induction on k Given a connected instance G → Check diamond-freeness → If it holds, construct G* → G = Δk(G’) ↔ G*=Δk+1(G’ V 2K1) Proof: Lemma: G is Kk+2-free ↔ Δk(G) is diamond-free ↔ each maximal clique of Δk(G) is either an isolated vertex or a (k+1)-clique, and any two maximal cliques intersect in at most one vertex. G1 G2

slide-33
SLIDE 33

Recognizing -anti-Gallai graphs

Induction on k Given a connected instance G → Check diamond-freeness → If it holds, construct G* → G = Δk(G’) ↔ G*=Δk+1(G’ V 2K1) Proof: Lemma: G is Kk+2-free ↔ Δk(G) is diamond-free ↔ each maximal clique of Δk(G) is either an isolated vertex or a (k+1)-clique, and any two maximal cliques intersect in at most one vertex. G1 G2 G*

slide-34
SLIDE 34

Recognition problem of -Gallai graphs

OPEN for each ≥ 2 ∶ What is the time complexity of the recognition problem of

  • Gallai graphs?

Remark. Deciding whether a given graph is the -Gallai graph of a -free graph → NP-complete (for each ≥ 3) But we have no polynomial-time checkable property P with: Γ(’) has property P if and only if ’ is -free

  • Gallai graph Γ()

vertices of Γ() ↔ -subgraphs of edges of Γ() ↔ two -subgraphs share a and NOT contained in a common in

slide-35
SLIDE 35

Thank you for your attention...