10/8/2015

Ma/CS 6a
Class 6: Introduction to Graphs

By Adam Sheffer

Today’s Class is about

Six Degrees of Kevin Bacon

1
 10/8/2015

Six Degrees of Kevin Bacon

 Claim. Any actor can be linked through
his/her film roles to Kevin Bacon within six
steps.
 Example. Keanu Reeves:

Graphs
Undirected Directed
graph graph

𝑎 𝑏 𝑎 𝑏

Edge
Vertex

2
 10/8/2015

Graph Representation
 We write 𝐺 = 𝑉, 𝐸 . That is, the graph 𝐺
has vertex set 𝑉 and edge set 𝐸.

 Example. In the figure:

◦ 𝑉 = 𝑎, 𝑏, 𝑐, 𝑑, 𝑒 . 𝑐 𝑑
◦ 𝐸 = { 𝑎, 𝑏 , 𝑎, 𝑑 , 𝑎, 𝑒 ,
𝑏, 𝑐 , 𝑏, 𝑒 }. 𝑎
𝑏

Graph Representation (cont.)

 𝑉 = 𝑎, 𝑏, 𝑐, 𝑑, 𝑒 .
 𝐸 = { 𝑎, 𝑏 , 𝑎, 𝑑 , 𝑎, 𝑒 , 𝑏, 𝑐 , 𝑏, 𝑒 }.

𝑐 𝑐
𝑎 𝑑

𝑒 𝑏 𝑎 𝑏

𝑑 𝑒

3
 10/8/2015

What is this Good For?

Cold war analysis Communication networks

Social networks Robot Motion

Planning

Paths and Cycles

𝑏 𝑎

Path Cycle
between 𝑎 through 𝑎
and 𝑏.
A cycle is a path that starts and ends in the same
vertex.

4
 10/8/2015

More on Paths and Cycles

 A path/cycle is said to be simple if it does
not visit any vertex more than once.

 The length of a path/cycle is the number

of edges that it consists of.

 Example. A simple cycle of

length 5.

Connectedness and Degrees

 A graph is connected if there is a path
between any two vertices.
 The degree of a vertex is the number of
edges that are adjacent to it.

Connected? Connected?

deg(𝑎) = 2 deg(𝑎) = 1

𝑎 𝑎

5
 10/8/2015

Simple Graphs
 An edge is a loop if both of its endpoints
are the same vertex.
 Two edges are parallel if they are
between the same pair of vertices.
 A graph is simple if it contains no loops
and no parallel edges.
 For now, we only consider simple graphs.

Parallel
A loop edges

Warm-up Exercise
 Prove. In any graph, the sum of the
degrees of the vertices is even.

 Proof. Every edge contributes 1 to the

degree of exactly two vertices. Thus,
deg 𝑣 = 2=2𝐸 .
𝑣∈𝑉 𝑒∈𝐸

6
 10/8/2015

Paths and Degrees

 Problem. Let 𝐺 = 𝑉, 𝐸 be a graph such
that the degree of every 𝑣 ∈ 𝑉 is at least
𝑑 (for some 𝑑 ≥ 2). Prove that 𝐺 contains
a path of length 𝑑.

A graph with minimum

degree 3.

Proof
 Assume for contradiction that a longest path 𝑃
is of length 𝑐 < 𝑑.
 Consider a vertex 𝑣 which is an endpoint of 𝑃.
 Since deg 𝑣 ≥ 𝑑 ≥ 𝑐 + 1, it must be
connected to at least one vertex 𝑢 ∉ 𝑃.
 By adding the edge 𝑣, 𝑢 to 𝑃, we obtain a
longer path, contradicting the maximality of 𝑃.

7
 10/8/2015

A Variant of the Problem

 Problem. Let 𝐺 = 𝑉, 𝐸 be a graph such
that the degree of every 𝑣 ∈ 𝑉 is at least
𝑑 (for some 𝑑 ≥ 2).
◦ Then there exists a cycle of length at least

𝑑+1

Connectivity Problem
 Prove. The vertices of a connected graph
𝐺 can always be ordered as 𝑣1 , 𝑣2 , … , 𝑣𝑛
such that for every 1 < 𝑖 ≤ 𝑛, if we
remove 𝑣𝑖 , 𝑣𝑖+1 , … , 𝑣𝑛 and the edges
adjacent to these vertices, 𝐺 remains
connected. 𝑣4
𝑣3

𝑣1 𝑣2

𝑣5

8
 10/8/2015

Proof
 Pick any vertex as 𝑣1 .
 Pick a vertex that is connected to 𝑣1 in 𝐺
and set it as 𝑣2 .
 Pick a vertex that is connected either to
𝑣1 or to 𝑣2 in 𝐺 and set it as 𝑣3 .
…

Back to Bacon
 We wish to build a graph for the problem.
 What are the vertices of the graph?
◦ A vertex for each actor.
 When is there an edge between two
vertices?
◦ When the corresponding actors played in a
common movie.
 Is the graph directed?
◦ No.

9
 10/8/2015

Example

 What should we do to check whether

every actor has a path to Kevin Bacon?
◦ Whether the graph is connected.

Bacon Numbers
 The Bacon number of an actor is the
minimum number of steps required to
connect him/her to Kevin Bacon.
 Example. By the picture below:
◦ Christopher Walken’s Bacon number is 2.
◦ Keanu Reeves’ Bacon number is 4.

10
 10/8/2015

Bacon Numbers (2)

 But Keanu Reeves’ Bacon number is 2!

 What should we do in the graph to find the correct

Bacon number of an actor?
◦ The length of the shortest path from the actor’s
vertex to Bacon’s vertex.

The BFS Algorithm

 We wish to find out:
◦ Whether the graph is connected.
◦ The shortest paths from Bacon’s vertex to
every other vertex.
 The BFS algorithm:
◦ Input. An undirected graph 𝐺 = (𝑉, 𝐸) and a
vertex 𝑠 ∈ 𝑉.
◦ Output. A shortest path from 𝑠 to any other
vertex of 𝐺 (if such a path exists).
◦ 𝐺 is connected if and only if there is a path to
every vertex.

11
 10/8/2015

Trees and Forests

 In an undirected graph, a tree is a
connected subgraph containing no cycles.
 A forest is a set of non-connected trees.

A forest with four trees

Rooted Trees
 A rooted tree is a tree with a special
vertex – the root – that is singled out.
 We draw the tree with the root on top,
and the edges “grow downwards”.
 A vertex 𝑣 is the parent of a vertex 𝑢 if
there is an edge 𝑢, 𝑣 and 𝑣 is above 𝑢.
◦ Each vertex, except for the root,
has a unique parent. 𝑠
𝑡

𝑠 is the root and 𝑡’s parent

12
 10/8/2015

BFS Output

𝑠
𝑠 𝑎 𝑏
𝑐 𝑎

𝑐 𝑑 𝑒 𝑑 𝑏

 The output is a BFS tree, containing only

shortest paths from 𝑠.
◦ A rooted tree with root 𝑠.

Test Your Intuition

 Problem. Is there always a unique tree
containing the shortest paths from 𝑠?

 Answer. No!

𝑠 𝑎
𝑠 𝑠

𝑏 𝑐 𝑏 𝑎 𝑏 𝑎
𝑐 𝑐

13
 10/8/2015

Erdős Numbers: the Math Version

 Paul Erdős (1913-1996). A Hungarian
mathematician. Possibly the most prolific
mathematician ever.
◦ People that wrote a paper with Erdős have an
Erdős number of 1.
◦ People that wrote a paper with someone that has
an Erdős number of 1, have an Erdős number of 2.
◦ Etc.
 Most leading scientists have a small
Erdős number.

BFS: Colors
 We call the vertex 𝑠 that we start from the
root of the tree.
 BFS scans the graph starting from the root.
 During the scan, every vertex has a color:
◦ Vertices that the algorithm did not visit yet are
colored white.
◦ Vertices that the algorithm did visit, but is not
yet done with are colored gray.
◦ Vertices that the algorithm is done with are
colored black.

14
 10/8/2015

Data Structure: Queue

 A queue stores “objects” (in our case –
vertices).
 Supports the operations:
◦ Enqueue – insert an object to the back of the
queue.
◦ Dequeue – remove an object from the front
of the queue.

BFS: The Main Idea

 A queue 𝑄 holds the vertices that are
currently gray. At first 𝑄 = 𝑠 .
 At each step, take out a vertex 𝑢 ∈ 𝑄 and
for every edge 𝑒 adjacent to 𝑢:
◦ If the other vertex of 𝑒 is gray or black, do
nothing.
◦ If the other vertex of 𝑒 is white, color it gray,
insert it into 𝑄, and insert 𝑒 to the BSF tree.
 After going over all of 𝑢’s edges, color 𝑢
black, and move to the next vertex in 𝑄.

15
 10/8/2015

Example: BFS Run

s s s s
f b f b f b f b
e e e e
c c c c
d d d d

s s s s
f b f b f b f b

e e e e
c c c c
d d d d

Example: Another BFS Run

16
 10/8/2015

The End

17