Scribd
Upload
7 views

LectureNote_MA221_13Nov

Uploaded by

telanganathletes2021
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
7 views

LectureNote_MA221_13Nov

Uploaded by

telanganathletes2021
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 20

Discrete Mathematics

Lecture 34: Graph Theory contd..


(Five Color Theorem and Connectedness)

Partha Sarathi Mandal


IIT Guwahati
Outline

• Proof of the five color theorem for planar graph

• Graph connectivity
• Vertex cut and vertex connectivity
• Edge cut and vertex connectivity
• Relation between vertex and edge connectivity
Five Color Theorem
• Theorem: Every planar graph can be 5 colored.
• Proof.
• We assume the graph has at least 5 vertices
• Else, the theorem will immediately follow.
• Next, in a planar graph, we see that there must be a vertex with
degree at most 5.
• Else, 2𝐸 = 𝑡𝑜𝑡𝑎𝑙 𝑑𝑒𝑔𝑟𝑒𝑒 ≥ 6𝑉, which contradicts with the fact
E ≤ 3𝑉 − 6
Five Color Theorem
• Theorem: Every planar graph can be 5 colored.
• Proof (contd.):
• Let 𝑣 be a vertex whose degree is at most 5
• Now, assume inductively that all planar graphs with 𝑛 − 1 vertices
can be colored in 5 colors  Thus if 𝑣 is removed, we can color the
graph properly in 5 colors
• What if we add back 𝑣 to the graph now?
• Case 1: Neighbors of 𝑣 uses at most 4 colors
• Case 2: Neighbors of 𝑣 uses up all 5 colors
Five Color Theorem
• Theorem: Every planar graph can be 5 colored.
• Proof (contd).:
• Case 1: Neighbors of 𝑣 uses at most 4 colors
• There is a 5th color for 𝑣
Five Color Theorem
• Theorem: Every planar graph can be 5 colored.
• Proof (contd).:
• Case 2: Neighbors of 𝑣 uses up all 5 colors

• Can we save 1 color, by coloring the


yellow neighbour in blue?
Five Color Theorem
• Theorem: Every planar graph can be 5 colored.
• Proof (Case 2, contd.):
• Can we color the yellow neighbour in blue?

• First, we check if the yellow


neighbour can connect to the
blue neighbour by a “switching”
yellow-blue path
Five Color Theorem
• Theorem: Every planar graph can be 5 colored.
• Proof (Case 2, contd.):
• Can we color the yellow neighbour in blue?
Five Color Theorem
• Theorem: Every planar graph can be 5 colored.
• Proof (Case 2, contd.):
• Can we color the yellow neighbour in blue?
Five Color Theorem
• Theorem: Every planar graph can be 5 colored.
• Proof (Case2, contd.):
• we color the orange neighbour in Green!

This means that there must be


two edges that cross each
other.

This contradicts the planarity


of the graph and hence it
concludes the proof.
Another illustration: Proof of Five Color Theorem
Case #2: deg(v) = 5. G-v can be colored with 5 colors.
If two of the neighbors of v are colored with the same color, Case #1: deg(v) ≤ 4.
then there is a color available for v. G-v can be colored with five colors

• This means that there must be two


edges that cross each other.
• This contradicts the planarity of
the graph and hence concludes the
proof.
Connectedness
Vertex Cut
• Vertex cut (separating set)
• Let 𝐺 = (𝑉, 𝐸) be a graph
• 𝑉′ ⊆ 𝑉 is a vertex cut, if G − 𝑉′ is disconnected
• Example: 𝑉 ′ = {𝑐, 𝑓}
𝑉 ′ = 𝑐, 𝑏, 𝑎
• Every connected graph with n nodes has a vertex cut?
• Yes, except the complete graph 𝐾𝑛
Vertex Connectivity
• Vertex connectivity of a graph 𝜅(𝐺)
• Size of the smallest vertex cut of 𝐺 𝜅=2
≈
• Minimum number of vertices to be deleted to disconnect the graph
• (Refine the definition) Minimum number of vertices to be deleted to either
disconnect the graph or produce a graph with a single node

0≤𝜅 𝐺 ≤𝑛−1
• 𝐺 is called 𝑘-connected, if 𝜅 𝐺 ≥ k

𝜅=2
Edge Cut
• Edge cut
• Let 𝐺 = (𝑉, 𝐸) be a graph
• 𝐸′ ⊆ 𝐸 is an edge cut, if G − 𝐸′ is disconnected
• Example: 𝐸 ′ = 𝑐, 𝑔 , 𝑐, 𝑓 , 𝑓, 𝑔 , 𝐸 ′ = 𝑎, 𝑑 , 𝑑, 𝑒 , 𝑑, 𝑐
• Every connected graph with n nodes has an edge cut?
• Yes, except for a graph with a single node
Edge Connectivity
• Edge connectivity of a graph 𝜆(𝐺)
𝜆=2
• Size of the smallest edge cut of 𝐺
≈
• Minimum number of edges to be deleted to disconnect the graph
• (Refine the definition) Minimum number of edges to be deleted to
either disconnect the graph or produce a graph with a single node
0≤𝜆 𝐺 ≤𝑛−1

𝜆=0
Upper Bound on Vertex Connectivity
• For any connected, non-complete graph
𝜅 𝐺 ≤ 𝑚𝑖𝑛𝑣 ∈ 𝑉 deg (𝑣)
• Consider the vertex 𝑣 with the least degree
• What happens if the neighbours of 𝑣 are deleted?
• The graph becomes disconnected
• Does there exit a connected, non-complete graph where
𝜅 𝐺 ≤ 𝑚𝑖𝑛𝑣 ∈ 𝑉 deg (𝑣)
Example to hold equality
min deg =2 (node d, e,..), 𝜅 = 1
Upper Bound on Edge Connectivity

• For any connected, non-complete graph


𝜆 𝐺 ≤ 𝑚𝑖𝑛𝑣 ∈ 𝑉 deg(𝑣)
• Consider the vertex 𝑣 with the least degree
• What happens if all the edges incident with 𝑣 are deleted?
• The graph becomes disconnected
• Does there exit a connected, non-complete graph where
𝜆 𝐺 ≤ 𝑚𝑖𝑛𝑣 ∈ 𝑉 deg (𝑣)

• Example to hold equality


• min deg =2 (node d, e,..), 𝜆 = 1
Relation Between Vertex and
Edge Connectivity
• Whitney’s Theorem: For any connected, non-complete
graph 𝜅 𝐺 ≤ 𝜆 𝐺
• Let E ′ = {𝑒1, . . , 𝑒𝜆} be a minimum edge cut of 𝐺, where
𝑒𝑖 = (𝑢𝑖, 𝑣𝑖)
• The nodes in {𝑢1, . . , 𝑢𝜆} /{𝑣1, . . , 𝑣𝜆} need not be distinct
• Delete u1, . . , u𝜆 − 1 from 𝐺 -- Graph 𝐻
• Edge e𝜆=(v𝜆, v𝜆) is a cut edge for 𝐻
• Else 𝐸′ is no a minimum edge cut for 𝐺
• Either u𝜆 OR v𝜆 is a cut vertex for 𝐻
• If u𝜆 is a cut vertex for 𝐻  u1, . . , u𝜆 − 1, u𝜆 is a vertex
cut for 𝐺
• If v𝜆 is a cut vertex for 𝐻  u1, . . , u𝜆 − 1, v𝜆 is a vertex
cut for 𝐺
Relation Between Vertex and Edge
Connectivity
• For any connected, non-complete graph 𝜅 𝐺 ≤ 𝜆 𝐺
• For any disconnected graph: 𝜅 𝐺 = 𝜆 𝐺 = 0
• For complete graphs 𝜅 𝐺 = 𝜆 𝐺 = 𝑛 − 1
• For any connected, non-complete graph:
𝜅 𝐺 ≤ 𝑚𝑖𝑛𝑣 ∈ 𝑉 deg(𝑣) 𝜆 𝐺 ≤ 𝑚𝑖𝑛𝑣 ∈ 𝑉 deg(𝑣)
• For any graph 𝐺:
𝜅 𝐺 ≤ 𝜆 𝐺 ≤ 𝑚𝑖𝑛𝑣 ∈ 𝑉 deg(𝑣)

You might also like