David Tankus Ariel University, ISRAEL Joint work with Vadim Levit - - PowerPoint PPT Presentation

david tankus
SMART_READER_LITE
LIVE PREVIEW

David Tankus Ariel University, ISRAEL Joint work with Vadim Levit - - PowerPoint PPT Presentation

Well-dominated graphs without cycles of lengths 4 and 5 David Tankus Ariel University, ISRAEL Joint work with Vadim Levit Ariel University, ISRAEL David Tankus Well-dominated graphs without cycles of lengths 4 and 5 Dominating sets A set S of


slide-1
SLIDE 1

Well-dominated graphs without cycles of lengths 4 and 5

David Tankus

Ariel University, ISRAEL

Joint work with

Vadim Levit Ariel University, ISRAEL

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-2
SLIDE 2

Dominating sets

A set S of vertices of G is dominating if every vertex of G is either in S or a neighbor of a vertex in S. A dominating set is minimal if it does not contain another dominating set. A dominating set is minimum if the graph does not admit a dominating set with a smaller cardinality.

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-3
SLIDE 3

Independent sets

An independent set is a set of pairwise non-adjacent vertices. An independent set is maximal if it is not contained in any larger independent set. An independent set is maximum if the graph does not admit an independent set with a higher cardinality.

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-4
SLIDE 4

Definitions

γ(G) = the cardinality of a minimum dominating set. Γ(G) = the maximal cardinality of a minimal dominating set. i(G) = the minimal cardinality of a maximal independent set. α(G) = the cardinality of a maximum independent set. γ(G) ≤ i(G) ≤ α(G) ≤ Γ(G)

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-5
SLIDE 5

Definitions

γ(G) = the cardinality of a minimum dominating set. Γ(G) = the maximal cardinality of a minimal dominating set. i(G) = the minimal cardinality of a maximal independent set. α(G) = the cardinality of a maximum independent set. γ(G) ≤ i(G) ≤ α(G) ≤ Γ(G)

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-6
SLIDE 6

Example

γ(G) ≤ i(G) ≤ α(G) ≤ Γ(G)

✉ ✉ ① ① ✉ ✁ ✁ ✁ ❆ ❆ ❆ ✑ ✑ ✑ ✑ ✑ ◗ ◗ ◗ ◗ ◗

❅ ❅ ❅ ❅ ❅ ✉ ✉ ① ✁ ✁ ✁ ❆ ❆ ❆ ✉ ✉ ✉ ✉ ✘✘✘✘✘ ✘ ❳❳❳❳❳ ❳ ❳❳❳❳❳ ❳ ✘✘✘✘✘ ✘

γ(G) = 3 i(G) = 4 α(G) = 6 Γ(G) = 7

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-7
SLIDE 7

Example

γ(G) ≤ i(G) ≤ α(G) ≤ Γ(G)

✉ ✉ ① ✉ ✉ ✁ ✁ ✁ ❆ ❆ ❆ ✑ ✑ ✑ ✑ ✑ ◗ ◗ ◗ ◗ ◗

❅ ❅ ❅ ❅ ❅ ✉ ✉ ① ✁ ✁ ✁ ❆ ❆ ❆ ① ① ✉ ✉ ✘✘✘✘✘ ✘ ❳❳❳❳❳ ❳ ❳❳❳❳❳ ❳ ✘✘✘✘✘ ✘

γ(G) = 3 i(G) = 4 α(G) = 6 Γ(G) = 7

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-8
SLIDE 8

Example

γ(G) ≤ i(G) ≤ α(G) ≤ Γ(G)

✉ ① ✉ ✉ ✉ ✁ ✁ ✁ ❆ ❆ ❆ ✑ ✑ ✑ ✑ ✑ ◗ ◗ ◗ ◗ ◗

❅ ❅ ❅ ❅ ❅ ✉ ✉ ① ✁ ✁ ✁ ❆ ❆ ❆ ① ① ① ① ✘✘✘✘✘ ✘ ❳❳❳❳❳ ❳ ❳❳❳❳❳ ❳ ✘✘✘✘✘ ✘

γ(G) = 3 i(G) = 4 α(G) = 6 Γ(G) = 7

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-9
SLIDE 9

Example

γ(G) ≤ i(G) ≤ α(G) ≤ Γ(G)

① ① ✉ ✉ ① ✁ ✁ ✁ ❆ ❆ ❆ ✑ ✑ ✑ ✑ ✑ ◗ ◗ ◗ ◗ ◗

❅ ❅ ❅ ❅ ❅ ✉ ✉ ✉ ✁ ✁ ✁ ❆ ❆ ❆ ① ① ① ① ✘✘✘✘✘ ✘ ❳❳❳❳❳ ❳ ❳❳❳❳❳ ❳ ✘✘✘✘✘ ✘

γ(G) = 3 i(G) = 4 α(G) = 6 Γ(G) = 7

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-10
SLIDE 10

Definitions

γ(G) ≤ i(G) ≤ α(G) ≤ Γ(G) i(G) = α(G) ⇐ ⇒ G is well-covered γ(G) = Γ(G) ⇐ ⇒ G is well-dominated Theorem (Finbow, Hartnell, Nowakowski, 1998) Every well-dominated graph is well-covered. There exist well-covered graphs which are not well-dominated:

✉ ✉ ✉ ✉ ✟✟✟✟✟ ✟ ✟✟✟✟✟ ✟ ✟✟✟✟✟ ✟ ✟✟✟✟✟ ✟ ✄ ✄ ✄ ✄ ✄✄ ✄ ✄ ✄ ✄ ✄✄ ✄ ✄ ✄ ✄ ✄✄ ✄ ✄ ✄ ✄ ✄✄ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ◗ ◗ ◗ ◗ ◗ ◗ ◗ ◗ ◗ ◗ ✉ ✉ ✉ ✉

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-11
SLIDE 11

Definitions

γ(G) ≤ i(G) ≤ α(G) ≤ Γ(G) i(G) = α(G) ⇐ ⇒ G is well-covered γ(G) = Γ(G) ⇐ ⇒ G is well-dominated Theorem (Finbow, Hartnell, Nowakowski, 1998) Every well-dominated graph is well-covered. There exist well-covered graphs which are not well-dominated:

✉ ✉ ✉ ✉ ✟✟✟✟✟ ✟ ✟✟✟✟✟ ✟ ✟✟✟✟✟ ✟ ✟✟✟✟✟ ✟ ✄ ✄ ✄ ✄ ✄✄ ✄ ✄ ✄ ✄ ✄✄ ✄ ✄ ✄ ✄ ✄✄ ✄ ✄ ✄ ✄ ✄✄ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ◗ ◗ ◗ ◗ ◗ ◗ ◗ ◗ ◗ ◗ ✉ ✉ ✉ ✉

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-12
SLIDE 12

Definitions

γ(G) ≤ i(G) ≤ α(G) ≤ Γ(G) i(G) = α(G) ⇐ ⇒ G is well-covered γ(G) = Γ(G) ⇐ ⇒ G is well-dominated Theorem (Finbow, Hartnell, Nowakowski, 1998) Every well-dominated graph is well-covered. There exist well-covered graphs which are not well-dominated:

✉ ✉ ✉ ✉ ✟✟✟✟✟ ✟ ✟✟✟✟✟ ✟ ✟✟✟✟✟ ✟ ✟✟✟✟✟ ✟ ✄ ✄ ✄ ✄ ✄✄ ✄ ✄ ✄ ✄ ✄✄ ✄ ✄ ✄ ✄ ✄✄ ✄ ✄ ✄ ✄ ✄✄ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ✪ ◗ ◗ ◗ ◗ ◗ ◗ ◗ ◗ ◗ ◗ ✉ ✉ ✉ ✉

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-13
SLIDE 13

Known Results

The complexity status of recognizing well-dominated graphs is not

  • known. It is even not known whether the problem is in NP.

Theorem (Finbow, Hartnell, Nowakowski, 1998) Recognizing well-dominated graphs with girth at least 6 can be done polynomially. Theorem (Finbow, Hartnell, Nowakowski, 1998) Recognizing well-dominated bipartite graphs can be done polynomially.

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-14
SLIDE 14

Known Results

The complexity status of recognizing well-dominated graphs is not

  • known. It is even not known whether the problem is in NP.

Theorem (Finbow, Hartnell, Nowakowski, 1998) Recognizing well-dominated graphs with girth at least 6 can be done polynomially. Theorem (Finbow, Hartnell, Nowakowski, 1998) Recognizing well-dominated bipartite graphs can be done polynomially.

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-15
SLIDE 15

Known Results

The complexity status of recognizing well-dominated graphs is not

  • known. It is even not known whether the problem is in NP.

Theorem (Finbow, Hartnell, Nowakowski, 1998) Recognizing well-dominated graphs with girth at least 6 can be done polynomially. Theorem (Finbow, Hartnell, Nowakowski, 1998) Recognizing well-dominated bipartite graphs can be done polynomially.

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-16
SLIDE 16

Definitions

Definition (Finbow, Hartnell, Nowakowski, 1994) A graph G is in the family F if there exists {x1, ..., xk} ⊆ V (G) such that xi is simplicial for each 1 ≤ i ≤ k, and {N [xi] : 1 ≤ i ≤ k} is a partition of V (G). The graph T10.

✉ ✉ ✉ ✉ ✉ ✉ ✉ ✉ ✉ ✉ ✟✟ ✟❍❍ ❍ ✟✟ ✟❍❍ ❍

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-17
SLIDE 17

Definitions

Definition (Finbow, Hartnell, Nowakowski, 1994) A graph G is in the family F if there exists {x1, ..., xk} ⊆ V (G) such that xi is simplicial for each 1 ≤ i ≤ k, and {N [xi] : 1 ≤ i ≤ k} is a partition of V (G). The graph T10.

✉ ✉ ✉ ✉ ✉ ✉ ✉ ✉ ✉ ✉ ✟✟ ✟❍❍ ❍ ✟✟ ✟❍❍ ❍

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-18
SLIDE 18

Definitions

Definition G( Ci1, ..., Cik) is the family of all graphs which do not contain cycles of lengths i1, ..., ik. The forbidden cycles are not necessarily induced. K10 ∈ G( C4)

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-19
SLIDE 19

Main result 1

Theorem (Finbow, Hartnell, Nowakowski, 1994) Let G ∈ G( C4, C5) be a connected graph. Then G is well-covered if and only if one of the following holds:

1

G is isomorphic to either C7 or T10.

2

G is a member of the family F. Theorem Let G ∈ G( C4, C5) be a connected graph. Then G is well-dominated if and only if it is well-covered.

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-20
SLIDE 20

Main result 1

Theorem (Finbow, Hartnell, Nowakowski, 1994) Let G ∈ G( C4, C5) be a connected graph. Then G is well-covered if and only if one of the following holds:

1

G is isomorphic to either C7 or T10.

2

G is a member of the family F. Theorem Let G ∈ G( C4, C5) be a connected graph. Then G is well-dominated if and only if it is well-covered.

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-21
SLIDE 21

Weighted graphs

Definition Let G be a graph and let w : V (G) − → R. Then mDSw(G) is the minimum weight of a dominating set. MDSw(G) is the maximum weight of a minimal dominating set. mISw(G) is the minimum weight of a maximal independent set. MISw(G) is the maximum weight of an independent set. mDSw(G) ≤ mISw(G) ≤ MISw(G) ≤ MDSw(G)

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-22
SLIDE 22

Weighted graphs

Definition Let G be a graph and let w : V (G) − → R. Then mDSw(G) is the minimum weight of a dominating set. MDSw(G) is the maximum weight of a minimal dominating set. mISw(G) is the minimum weight of a maximal independent set. MISw(G) is the maximum weight of an independent set. mDSw(G) ≤ mISw(G) ≤ MISw(G) ≤ MDSw(G)

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-23
SLIDE 23

Weighted graphs

mDSw(G) ≤ mISw(G) ≤ MISw(G) ≤ MDSw(G) Definition Let G be a graph and let w : V (G) − → R. Then G is w-well-covered if mISw(G) = MISw(G). G is w-well-dominated if mDSw(G) = MDSw(G). If G is w-well-dominated then G is w-well-covered.

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-24
SLIDE 24

Weighted graphs

mDSw(G) ≤ mISw(G) ≤ MISw(G) ≤ MDSw(G) Definition Let G be a graph and let w : V (G) − → R. Then G is w-well-covered if mISw(G) = MISw(G). G is w-well-dominated if mDSw(G) = MDSw(G). If G is w-well-dominated then G is w-well-covered.

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-25
SLIDE 25

Vector space

Theorem Let G be a graph. Then the set of weight functions w : V (G) − → R such that G is w-well-dominated is a vector space. Proof: ∀v ∈ V w(v) = w1(v) + λw2(v) w(S) =

  • s∈S

w(s) =

  • s∈S

(w1(s) + λw2(s)) =

  • s∈S

w1(s) + λ

  • s∈S

w2(s) = t1 + λt2

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-26
SLIDE 26

Vector space

Theorem Let G be a graph. Then the set of weight functions w : V (G) − → R such that G is w-well-dominated is a vector space. Proof: ∀v ∈ V w(v) = w1(v) + λw2(v) w(S) =

  • s∈S

w(s) =

  • s∈S

(w1(s) + λw2(s)) =

  • s∈S

w1(s) + λ

  • s∈S

w2(s) = t1 + λt2

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-27
SLIDE 27

Vector space

Definition The vector space of weight functions w : V (G) − → R such that G is w-well-covered is denoted by WCW (G). Definition The vector space of weight functions w : V (G) − → R such that G is w-well-dominated is denoted by WWD(G). Theorem WWD(G) is a subspace of WCW (G).

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-28
SLIDE 28

Vector space

Definition The vector space of weight functions w : V (G) − → R such that G is w-well-covered is denoted by WCW (G). Definition The vector space of weight functions w : V (G) − → R such that G is w-well-dominated is denoted by WWD(G). Theorem WWD(G) is a subspace of WCW (G).

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-29
SLIDE 29

Definitions

Definition L(G) is the set of all vertices v ∈ V (G) such that either d(v) = 1

  • r v is on a triangle and d(v) = 2.

① ❆ ❆ ❆ ❆ ❆ ✁ ✁ ✁ ✁ ✁ ① ❆ ❆ ❆ ❆ ❆ ✁ ✁ ✁ ✁ ✁ ✉ ✉ ✉ ✉ ✉ ✉ ✉ ① ① ✉ ✉ ① ① ✁ ✁ ✁ ✁ ✁ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ✉ ✉ ✉ ✉ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ① ✉ ✉ ① ① ① ❆ ❆ ❆ ❆ ❆ ✁ ✁ ✁ ✁ ✁ ① ①

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-30
SLIDE 30

Definitions

Definition D(v)= N(v) \ N(N2(v)) = N(v) ∩ L(G). M(v) is a maximal independent set of D(v).

① ❆ ❆ ❆ ❆ ❆ ✁ ✁ ✁ ✁ ✁ ① ❆ ❆ ❆ ❆ ❆ ✁ ✁ ✁ ✁ ✁ ✉ ✉ ✉ ✉ ✉ ✉ ✉ ① ① ✉ ✉ ① ① ✁ ✁ ✁ ✁ ✁ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ✉ ✉ ✉ ✉ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ① ✉ ✉ ① ① ① ❆ ❆ ❆ ❆ ❆ ✁ ✁ ✁ ✁ ✁ ① ①

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-31
SLIDE 31

WCW (G) when G ∈ G( C4, C5, C6)

Theorem (Levit, Tankus, 2015) Let G ∈ G( C4, C5, C6) be a graph, and let w : V (G) − → R. Then G is w-well-covered if and only if one of the following holds:

1

G is isomorphic to either C7 or T10, and there exists a constant k ∈ R such that w ≡ k.

2

The following conditions hold:

G is isomorphic to neither C7 nor T10. For every two vertices, l1 and l2, in the same component of L(G) it holds that w(l1) = w(l2). For every v ∈ V (G) \ L(G) it holds that w(v) = w(M(v)) for some maximal independent set M(v) of D(v).

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-32
SLIDE 32

WWD(G) WCW (G)

✉ ✉ ① ① ①

a b c a + b b + c

❅ ❅ ❅ ❅ ❅ ❅

  • w ∈ WCW (G)

w ∈ WWD(G) ⇐ ⇒ b = 0

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-33
SLIDE 33

WWD(G) WCW (G)

✉ ✉ ① ① ①

a b c a + b b + c

❅ ❅ ❅ ❅ ❅ ❅

  • w ∈ WCW (G)

w ∈ WWD(G) ⇐ ⇒ b = 0

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-34
SLIDE 34

WWD(G) WCW (G)

✉ ✉ ① ① ①

a b c a + b b + c

❅ ❅ ❅ ❅ ❅ ❅

  • w ∈ WCW (G)

w ∈ WWD(G) ⇐ ⇒ b = 0

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-35
SLIDE 35

L∗(G)

v v2 v1

❆ ❆ ❆ ❆ ❆ ✁ ✁ ✁ ✁ ✁ ✉ ❆ ❆ ❆ ❆ ❆ ✁ ✁ ✁ ✁ ✁ ① ✉ ✉ ✉ ✉ ✉ ✉ ① ✉ ✉ ✉ ① ① ✁ ✁ ✁ ✁ ✁ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ✉ ✉ ✉ ✉ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ① ① ✉ ① ① ① ❆ ❆ ❆ ❆ ❆ ✁ ✁ ✁ ✁ ✁ ① ✉

S is a maximal independent set of G \ N2[v]. S dominates neither N(v1) ∩ N2(v) nor N(v2) ∩ N2(v).

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-36
SLIDE 36

L∗(G)

v v2 v1

❆ ❆ ❆ ❆ ❆ ✁ ✁ ✁ ✁ ✁ ① ❆ ❆ ❆ ❆ ❆ ✁ ✁ ✁ ✁ ✁ ① ✉ ① ✉ ✉ ① ✉ ① ✉ ✉ ✉ ① ① ✁ ✁ ✁ ✁ ✁ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ① ✉ ✉ ✉ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ① ① ✉ ① ① ① ❆ ❆ ❆ ❆ ❆ ✁ ✁ ✁ ✁ ✁ ① ✉

S is a maximal independent set of G \ N2[v]. S1 is a maximal independent set of (N(v1) ∩ N2(v)) \ N(S). S2 is a maximal independent set of (N(v2) ∩ N2(v)) \ N(S).

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-37
SLIDE 37

L∗(G)

v v2 v1

❆ ❆ ❆ ❆ ❆ ✁ ✁ ✁ ✁ ✁ ① ❆ ❆ ❆ ❆ ❆ ✁ ✁ ✁ ✁ ✁ ① ✉ ① ✉ ✉ ① ✉ ① ✉ ✉ ✉ ① ① ✁ ✁ ✁ ✁ ✁ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ① ✉ ✉ ✉ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ① ① ✉ ① ① ① ❆ ❆ ❆ ❆ ❆ ✁ ✁ ✁ ✁ ✁ ① ✉

T = S ∪ S1 ∪ S2 ∪ {v} S ∪ S2 ∪ {v1} = ⇒ w(v1) = w(v) + w(S1) S ∪ S1 ∪ {v2} = ⇒ w(v2) = w(v) + w(S2) S ∪ {v1, v2} = ⇒ w(v1) + w(v2) = w(v) + w(S1) + w(S2)

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-38
SLIDE 38

L∗(G)

v v2 v1

❆ ❆ ❆ ❆ ❆ ✁ ✁ ✁ ✁ ✁ ① ❆ ❆ ❆ ❆ ❆ ✁ ✁ ✁ ✁ ✁ ① ✉ ① ✉ ✉ ① ✉ ① ✉ ✉ ✉ ① ① ✁ ✁ ✁ ✁ ✁ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ① ✉ ✉ ✉ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ① ① ✉ ① ① ① ❆ ❆ ❆ ❆ ❆ ✁ ✁ ✁ ✁ ✁ ① ✉

T = S ∪ S1 ∪ S2 ∪ {v} S ∪ S2 ∪ {v1} = ⇒ w(v1) = w(v) + w(S1) S ∪ S1 ∪ {v2} = ⇒ w(v2) = w(v) + w(S2) S ∪ {v1, v2} = ⇒ w(v1) + w(v2) = w(v) + w(S1) + w(S2)

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-39
SLIDE 39

L∗(G)

v v2 v1

❆ ❆ ❆ ❆ ❆ ✁ ✁ ✁ ✁ ✁ ① ❆ ❆ ❆ ❆ ❆ ✁ ✁ ✁ ✁ ✁ ① ✉ ① ✉ ✉ ① ✉ ① ✉ ✉ ✉ ① ① ✁ ✁ ✁ ✁ ✁ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ① ✉ ✉ ✉ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ① ① ✉ ① ① ① ❆ ❆ ❆ ❆ ❆ ✁ ✁ ✁ ✁ ✁ ① ✉

T = S ∪ S1 ∪ S2 ∪ {v} S ∪ S2 ∪ {v1} = ⇒ w(v1) = w(v) + w(S1) S ∪ S1 ∪ {v2} = ⇒ w(v2) = w(v) + w(S2) S ∪ {v1, v2} = ⇒ w(v1) + w(v2) = w(v) + w(S1) + w(S2)

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-40
SLIDE 40

L∗(G)

v v2 v1

❆ ❆ ❆ ❆ ❆ ✁ ✁ ✁ ✁ ✁ ① ❆ ❆ ❆ ❆ ❆ ✁ ✁ ✁ ✁ ✁ ① ✉ ① ✉ ✉ ① ✉ ① ✉ ✉ ✉ ① ① ✁ ✁ ✁ ✁ ✁ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ① ✉ ✉ ✉ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ① ① ✉ ① ① ① ❆ ❆ ❆ ❆ ❆ ✁ ✁ ✁ ✁ ✁ ① ✉

T = S ∪ S1 ∪ S2 ∪ {v} S ∪ S2 ∪ {v1} = ⇒ w(v1) = w(v) + w(S1) S ∪ S1 ∪ {v2} = ⇒ w(v2) = w(v) + w(S2) S ∪ {v1, v2} = ⇒ w(v1) + w(v2) = w(v) + w(S1) + w(S2)

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-41
SLIDE 41

L∗(G)

v v2 v1

❆ ❆ ❆ ❆ ❆ ✁ ✁ ✁ ✁ ✁ ① ❆ ❆ ❆ ❆ ❆ ✁ ✁ ✁ ✁ ✁ ① ✉ ① ✉ ✉ ① ✉ ① ✉ ✉ ✉ ① ① ✁ ✁ ✁ ✁ ✁ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ① ✉ ✉ ✉ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ① ① ✉ ① ① ① ❆ ❆ ❆ ❆ ❆ ✁ ✁ ✁ ✁ ✁ ① ✉

w(v) = 0

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-42
SLIDE 42

L∗(G)

Definition L∗(G) is the set of all vertices v ∈ V (G) such that one of the following holds: d(v) = 1. The following conditions hold:

d(v) = 2. v is on a triangle, (v, v1, v2). Every maximal independent set of V (G) \ N2[v] dominates at least one of N(v1) ∩ N2(v) and N(v2) ∩ N2(v).

L∗(G) ⊆ L(G). v ∈ L(G) \ L∗(G) if and only if the following conditions hold: d(v) = 2 v is on a triangle, (v, v1, v2). There exists a maximal independent set of V (G) \ N2[v] which dominates neither N(v1) ∩ N2(v) nor N(v2) ∩ N2(v).

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-43
SLIDE 43

L∗(G)

Definition L∗(G) is the set of all vertices v ∈ V (G) such that one of the following holds: d(v) = 1. The following conditions hold:

d(v) = 2. v is on a triangle, (v, v1, v2). Every maximal independent set of V (G) \ N2[v] dominates at least one of N(v1) ∩ N2(v) and N(v2) ∩ N2(v).

L∗(G) ⊆ L(G). v ∈ L(G) \ L∗(G) if and only if the following conditions hold: d(v) = 2 v is on a triangle, (v, v1, v2). There exists a maximal independent set of V (G) \ N2[v] which dominates neither N(v1) ∩ N2(v) nor N(v2) ∩ N2(v).

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-44
SLIDE 44

L∗(G)

Definition L∗(G) is the set of all vertices v ∈ V (G) such that one of the following holds: d(v) = 1. The following conditions hold:

d(v) = 2. v is on a triangle, (v, v1, v2). Every maximal independent set of V (G) \ N2[v] dominates at least one of N(v1) ∩ N2(v) and N(v2) ∩ N2(v).

L∗(G) ⊆ L(G). v ∈ L(G) \ L∗(G) if and only if the following conditions hold: d(v) = 2 v is on a triangle, (v, v1, v2). There exists a maximal independent set of V (G) \ N2[v] which dominates neither N(v1) ∩ N2(v) nor N(v2) ∩ N2(v).

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-45
SLIDE 45

L∗(G)

x x2 x1 y y2 y1

❆ ❆ ❆ ❆ ❆ ✁ ✁ ✁ ✁ ✁ ① ❆ ❆ ❆ ❆ ❆ ✁ ✁ ✁ ✁ ✁ ✉ ✉ ✉ ✉ ✉ ✉ ✉ ① ① ✉ ✉ ① ① ✁ ✁ ✁ ✁ ✁ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ✉ ✉ ✉ ✉ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ① ✉ ✉ ① ① ① ❆ ❆ ❆ ❆ ❆ ✁ ✁ ✁ ✁ ✁ ① ①

L(G) \ L∗(G) = {x, y}

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-46
SLIDE 46

WWD(G) when G ∈ G( C4, C5, C6)

Theorem Let G ∈ G( C4, C5, C6) be a connected graph, and let w : V (G) − → R. Then G is w-well-dominated if and only if one of the following holds:

1

G is isomorphic to either C7 or T10, and there exists a constant k ∈ R such that w ≡ k.

2

The following conditions hold:

1

G is isomorphic to neither C7 nor T10.

2

For every two vertices, l1 and l2, in the same component of L(G) it holds that w(l1) = w(l2).

3

w(v) = 0 for every vertex v ∈ L(G) \ L∗(G).

4

For every v ∈ V (G) \ L(G) it holds that w(v) = w(M(v)) for some maximal independent set M(v) of D(v).

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-47
SLIDE 47

Proof

Let N(L∗(G)) = {v1, ..., vk}. For each 1 ≤ i ≤ k let Ti = {vi} ∪ D(vi). S ∩ Ti ∈ {{vi}, M(vi)} = ⇒ w(S ∩ Ti) = w(vi).

① ❆ ❆ ❆ ❆ ❆ ✁ ✁ ✁ ✁ ✁ ① ❆ ❆ ❆ ❆ ❆ ✁ ✁ ✁ ✁ ✁ ✉ ✉ ✉ ✉ ① ① ✉ ① ① ✁ ✁ ✁ ✁ ✁ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ❆ ✉ ① ① ① ① ① ① ① ✁ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ✁ ① ✉ ✉ ① ① ① ❆ ❆ ❆ ❆ ❆ ✁ ✁ ✁ ✁ ✁ ① ①

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-48
SLIDE 48

Proof

w(S) = w(S \ (

  • 1≤i≤k

Ti)) +

  • 1≤i≤k

w(S ∩ Ti) −

  • 1≤i<j≤k

w(S ∩ Ti ∩ Tj) = = 0 +

  • 1≤i≤k

w(vi) − 0 =

  • 1≤i≤k

w(vi). mDSw(G) = MDSw(G) =

  • 1≤i≤k

w(vi)

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-49
SLIDE 49

Proof

w(S) = w(S \ (

  • 1≤i≤k

Ti)) +

  • 1≤i≤k

w(S ∩ Ti) −

  • 1≤i<j≤k

w(S ∩ Ti ∩ Tj) = = 0 +

  • 1≤i≤k

w(vi) − 0 =

  • 1≤i≤k

w(vi). mDSw(G) = MDSw(G) =

  • 1≤i≤k

w(vi)

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-50
SLIDE 50

Proof

w(S) = w(S \ (

  • 1≤i≤k

Ti)) +

  • 1≤i≤k

w(S ∩ Ti) −

  • 1≤i<j≤k

w(S ∩ Ti ∩ Tj) = = 0 +

  • 1≤i≤k

w(vi) − 0 =

  • 1≤i≤k

w(vi). mDSw(G) = MDSw(G) =

  • 1≤i≤k

w(vi)

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-51
SLIDE 51

dim(WWD(G))

Corollary Assume G ∈ G( C4, C5, C6). Then dim(WWD(G)) = α(G[L∗(G)]) L∗(G) = L(G) ⇐ ⇒ WWD(G) = WCW (G).

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-52
SLIDE 52

Future Work

Problem Discover more cases, where recognizing well-dominated graphs and/or finding WWD(G) can be done polynomially. Problem Characterize all graphs, where the equality WCW (G) = WWD(G) holds.

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-53
SLIDE 53

Future Work

Problem Discover more cases, where recognizing well-dominated graphs and/or finding WWD(G) can be done polynomially. Problem Characterize all graphs, where the equality WCW (G) = WWD(G) holds.

David Tankus Well-dominated graphs without cycles of lengths 4 and 5

slide-54
SLIDE 54

The end

Thank you

David Tankus Well-dominated graphs without cycles of lengths 4 and 5