Graph Theory: Intro and Trees

CS 2800: Discrete Structures, Spring 2015

Sid Chaudhuri
 This is ok

(Ok because equality is symmetric and transitive)

implied if you omit them
This is NOT ok

⇒
These ⇒ symbols are

⇒
⇒
⇒
⇒
… which is true, so QED No!
 Plea for the Day #1

Please read out your proofs in plain

English and ask yourself if it makes sense

http://www.plainenglish.co.uk/
 Plea for the Day #2

Plain English is often better than

“mathy” notation

Instead of “∀d∈Days,Rainy(d)⇒Umbrella(d)”, say

“If it's a rainy day, I will carry an umbrella”

(Which is easier to read and debug?)

(But do be precise and terse)
(This is an acquired skill – look at lots of well-written proofs)
<modest>... such as the homework and prelim solutions</modest>
These aren't the graphs we're interested in
This is

V.J. Wedeen and L.L. Wald, Martinos Center for Biomedical Imaging at MGH
And so is this
A social graph
An older social graph
A fictional social graph
A transport graph
Another transport graph
A circuit graph (flip-flop)
A circuit graph (Intel 4004)
A circuit graph (Intel Haswell)
 A probabilistic graphical model
P(C=F) P(C=T)
0.5 0.5

C P(S=F) P(S=T) Cloudy C P(R=F) P(R=T)

F 0.5 0.5 F 0.8 0.2
T 0.9 0.1 T 0.2 0.8

Sprinkler Rain

P(W, R, S, C) =
P(W | R, S) Wet grass R S P(W=F) P(W=T)

× P(R | C) F F 1.0 0.0

T F 0.1 0.9
× P(S | C) F T 0.1 0.9
× P(C) T T 0.01 0.99
This is a graph(ical model) that has learned
to recognize cats
Some abstract graphs

K5
K3,3

=
 What is a graph?
A
● An undirected graph G = (V, E)
consists of D
C
– A non-empty set of vertices/nodes V B
– A set of edges E, each edge being a set of one or two vertices (if
one vertex, the edge is a self-loop)
A
● A directed graph G = (V, E)
consists of D
C
– A non-empty set of vertices/nodes V B

– A set of edges E, each edge being an ordered pair of vertices (the

first vertex is the “start” of the edge, the second is the “end” )
● That is, E ⊆ V × V, or E is a relation from V to V
 Multigraphs

D
C

● Multiple edges between same pair of vertices

● Need a different representation (not a relation)
 Adjacency, Incidence and Degree
● Two vertices are adjacent iff there is an edge between
them
● An edge is incident on both of its vertices
● Undirected graph:
– Degree of a vertex is the number of edges incident on it
● Directed graph:
– Outdegree of a vertex u is the number of edges leaving it,
i.e. the number of edges (u, v)
– Indegree of a vertex u is the number of edges entering it,
i.e. the number of edges (v, u)
 Paths
● A path is a sequence of vertices v0, v1, v2 … vn, all different except
possibly the first and the last, such that
– (in an undirected graph) every pair {vi, vi + 1} is an edge
– (in a directed graph) every pair (vi, vi + 1) is an edge
● Alternatively, a path may be defined as a sequence of distinct
edges e0, e1, e2 … en such that
– Every pair ei, ei + 1 shares a vertex
– These vertices are distinct, except possibly the first & the last
– If the graph is directed, then the end vertex of ei is the start vertex
of ei + 1 (the “arrows” point in a consistent direction)
● We will use these interchangeably
● A cycle is a path with the same first and last vertex
 Connectedness
● An undirected graph is connected iff for every pair of
vertices, there is a path containing them
● A directed graph is strongly connected iff it satisfies the
above condition for all ordered pairs of vertices (for
every u, v, there are paths from u to v and v to u)
● A directed graph is weakly connected iff replacing all
directed edges with undirected ones makes it connected

A A

C D C D

B B
Connected Not connected
 Trees
● A forest is an undirected graph with no cycles
A
F G

C D
A B C D E
B

● A tree is a connected forest (← definition)

G
A

E G
C D

B
A B C D E
 Identifying trees
● An undirected graph G on a finite set of vertices is a
tree iff any two of the following conditions hold

– |E| = |V| - 1

– G is connected
– G has no cycles

● Any two of these imply the third

● There are many other elegant characterizations of
trees
 Proof Sketches
● If G is connected and lacks cycles, then |E| = |V| - 1
– Show that such a graph always has a vertex of degree 1
– Use induction, repeatedly removing such a vertex
● If G is connected and |E| = |V| - 1, then it lacks cycles
– Show that a connected graph has a spanning tree
– Apply the |E| = |V| - 1 formula to the spanning tree
● If G lacks cycles and |E| = |V| - 1, then it is connected
– If disconnected, must have ≥ 2 connected components, each
of which must be a tree
– Sum vertex and edge counts over connected components,
show that they don't add up
 Cayley's Formula
● For any positive integer n, the number of trees on
n labeled vertices is nn – 2

2 vertices: 1 tree 3 vertices: 3 trees

if ul
e au t 4 vertices: 16 trees

n y b
Ma roofs!
p