Scribd
Upload
0 views

1comb-lecture4-basicgraph1

The document introduces the concept of graphs in discrete mathematics, defining a graph as an ordered pair consisting of a set of vertices and edges. It discusses various types of graphs, including multigraphs and directed graphs, and explains key concepts such as isomorphism, connectedness, and the degree of vertices. Additionally, it covers graph presentations, adjacency matrices, and the conditions for a sequence to be a graph score.

Uploaded by

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

1comb-lecture4-basicgraph1

The document introduces the concept of graphs in discrete mathematics, defining a graph as an ordered pair consisting of a set of vertices and edges. It discusses various types of graphs, including multigraphs and directed graphs, and explains key concepts such as isomorphism, connectedness, and the degree of vertices. Additionally, it covers graph presentations, adjacency matrices, and the conditions for a sequence to be a graph score.

Uploaded by

allrounderguno
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 24

4 Graphs: an introduction

Many situations in various practically motivated problems and also in mathematics and
theoretical computer science can be captured by a scheme consisting of two things:

- a (finite) set of points,


- lines joining some pairs of the points.

Example)

Point Line
Participants at a birthday party pairs of participants who know each other
Street crossings in a city Streets
Subway stations Railways
Websites Links between websites
If we disregard the length, shape, and other properties of the joins, and
we only pay attention to which pairs of points are joint,
then we wrraive at the mathematical notion of a graph.

Although very simple, a graph is one of the key concepts in discrete mathematics.

[Definition 4.1.1]
A graph G is an ordered pair (V, E) where
V is a set and E is a set of subsets of V of size 2.

The elements of V are called vertices and


the elements of E are called edges.
If {u, v} is an edge, we simply write uv.
We also say that u is adjacent to v or u is a neighbor of v.

For a vertex v, NG(v) denotes the set of neighbors of v in G.


The size of NG(v) is called the degree of v in G.

Let G and H be graphs.


We say that H is a subgraph of G if V(H) ⊆ V(G) and E(H) ⊆ E(G).
We say that H is an induced subgraph of G

( 2 )
V(H)
if V(H) ⊆ V(G) and E(H) = E(G) ∩

Graphs are finite / infinite, countable … according to their order.


In this lecture, graphs will be all finite.
Some variants :

Multigraphs - allow multiple edges between two vertices


allow loops (ends of an edge are the same)

Directed graphs - edges are defined as ordered pairs (a, b)

(We will sometimes use them.)


Some important graphs. Let [n] = {1,2,…, n}.

( 2 )
V(Kn)
Kn : V(Kn) = [n] and E(Kn) = .

Km,n : V(Km,n) = {ai : i ∈ [m]} ∪ {bi : i ∈ [n]} and


E(Kn,n) = {aibj : (i, j) ∈ [m] × [n]}

Pn : V(Pn) = [n] and E(Pn) = {{i, i + 1} : i ∈ [n − 1]}

Cn : V(Cn) = [n] and E(Cn) = {{i, i + 1} : i ∈ [n − 1]} ∪ {{n,1}}

Normally, any graph isomorphic to Kn is simply written as Kn .. and so on.


Let G and H be two graphs.

We say that G is isomorphic to H, written as G ≅ H,


if there is a bijection ϕ : V(G) → V(H) such that
- xy ∈ E(G) if and only if ϕ(x)ϕ(y) ∈ E(H).

Such a mapping is called an isomorphism.

If G = H, then it is called an automorphism.


[Question] How many pairwise non-isomorphic graphs on 3 vertices?
[Question] How many pairwise non-isomorphic graphs on 3 vertices?

[Question] How many pairwise non-isomorphic graphs on n vertices?


For two graphs G and H, we write
G ∪ H = (V(G) ∪ V(H), E(G) ∪ E(H)) and
G ∩ H = (V(G) ∩ V(H), E(G) ∩ E(H)).

If G ∩ H = ∅, then G and H are disjoint (sometimes, vertex-disjoint).

For U ⊆ V(G), we denote by G[U] the induced subgraph of G on U.


For U ⊆ V(G),
G − U is the graph obtained by removing vertices in U, that is,
G − U = G[V(G)∖U].
We write G − v = G − {v} for v ∈ V(G).

( 2 )
V(G)
For F ⊆ ,

G − F is the graph obtained by removing edges in F, that is


G − F = (V(G), E(G)∖F), and
G + F is the graph obtained by adding edges in F,
G + F = (V(G), E(G) ∪ F).

We write G − e and G + e for a single edge e.


A walk of length k in a graph G is a non-empty alternating sequence
(v0, e0, v1, e1, ⋯ek−1, vk) of vertices and edges such that
ei = vivi+1 for all i ∈ {0,…, k − 1}.

If v0 = vk, then the walk is closed.


If the vertices are all distinct, then it is called a path.

Let P be a path (v0, e0, v1, e1, ⋯ek−1, vk).


v0 and vk are called endvertices or ends of P, and other vertices are internal vertices.
We often refer to a path by the natural sequence of vertices;
P = v0v1⋯vk. We say P is a path between v0 and vk.
[Exercise] Let G be a graph and v, w ∈ V(G).
Show that every walk from x to y contains a path from x to y.
A cycle in a graph is a non-empty subgraph C = (V, E) of the form
V = {x0, x1, …, xk−1} and E = {x0x1, x1x2, …, xk−2xk−1, xk−1x0},
where the xi are all distinct.

The number of edges is the length of C.


A graph G is connected if it is non-empty and
any two vertices of G are linked by a path.

A graph is disconnected if it is not connected.

Let G be a graph.
A maximal connected subgraph of G is a ‘connected component’ of G (or component).
The components are induced subgraphs, and their vertex sets partition V(G).

(One can also define as induced subgraphs on equivalent classes where


equivalence relation x ∼ y iff x and y is connected by a path. )
The distance distG(x, y) between x and y
is the length of a shortest path between x and y in G.
If there is no path, we define it to be ∞.

Distance function has the following properties:


1. distG(v, v′) ≥ 0 and distG(v, v′) = 0 iff v = v′
2. (symmetry) distG(v, v′) = distG(v′, v)
3. (triangle inequality) distG(v, z) ≤ distG(v, w) + distG(w, z) for any v, w, z ∈ V(G).

(Usually, a space with such a function is called a metric space)







Graph presentations

Let v1, v2, …, vn be the vertices of a graph G.

n
The adjacency matrix of G is an (n × n) matrix AG = (aij)i,j=1 such that

aij = 1 if vivj ∈ E(G) and aij = 0 otherwise.

Analyzing graphs using matrix theory is called ‘algebraic graph theory’.


But not only for algebraic graph theory, adjacency matrices are broadly used.
4.3 Graph score

Let G be a graph and v1, v2, …, vn be vertices of G.

The sequence (degG(v1), degG(v2), …, degG(vn)) is called


a degree sequence of G, or a score of G.

We will not distinguish two scores if one of them can be obtained from the other
by rearranging the order of the numbers.

We will usually write scores in nondecreasing order.

NOTE: Graphs with the same score might be non-isomorphic. Example?


[Question] A non-decreasing sequence of number (a1, a2, …, an) is given.
When is it a score of some graph?
Can you find some way to determine?
[Question] A non-decreasing sequence of number (a1, a2, …, an) is given.
When is it a score of some graph?
Can you find some way to determine?

Some basic condition:


each ai ≤ n − 1.

∑
[Proposition 4.3.1] For every graph G, degG(v) = 2 | E(G) | .
v∈V(G)
[Question] A non-decreasing sequence of number (a1, a2, …, an) is given.
When is it a score of some graph?
Can you find some way to determine?

Some basic condition:


each ai ≤ n − 1.

∑
[Proposition 4.3.1] For every graph G, degG(v) = 2 | E(G) | .
v∈V(G)

Full characterization is not quite simple, and it is related to network flows.


We present a simple algorithm for deciding whether a given sequence is a score or not.
[Theorem 4.3.3 (Score theorem)]
Let D = (d1, d2, …, dn) be a non-decreasing sequence of natural numbers with n > 1.
Let D′ = (d′1, d′2, …, d′n−1) such that
d′i = di for i < n − dn and
d′i = di − 1 for i ≥ n − dn.

Then D is a graph score if and only if D′ is a graph score.

Ex) D = (1,1,2,2,2,3,3) and D′ = (1,1,2,1,1,2).










PF) (<=) Easy direction.
Suppose that D′ is a score of some graph G′ where V(G′) = {v1, …, vn−1} and
degG′(vi) = d′i for all i ∈ [n − 1].

Fix a new vertex vn distinct from v1, …, vn−1 and define a new graph G where
V(G) = V(G′) ∪ {vn}
E(G) = E(G′) ∪ {vivn : i = n − dn, n − dn + 1,…, n − 1}

(So, vn is connected to the last dn vertices of G′)

Clearly, G has score D.

It is more difficult to prove the reverse.










PF) (=>) Let be the set of all graphs on the vertex set {v1, …, vn} in which
the degree of each vertex vi is di.

(Claim) contains a graph G0 where vn is adjacent to exactly vn−dn, vn−dn+1, …, vn−1.

If dn = n − 1, then we are done. We may assume that dn < n − 1.


For G ∈ , let j(G) be the largest index j ∈ [n − 1] where vjvn ∉ E(G).

Let G0 ∈ be a graph with minimum j(G0).


We prove that j(G0) = n − dn − 1, which implies the claim.
𝒢
𝒢
𝒢
𝒢
PF) (=>) We prove that j(G0) = n − dn − 1, which implies the claim.

Suppose that j(G0) > n − dn − 1.


The vertex vn has to be adjacent to dn vertices, and
at most dn − 1 of them can have a large index than vj(G0).

So, there exists i < j(G0) such that vivn ∈ E(G).


Since degG0(vi) ≤ degG0(vj), there exists a vertex vk adjacent to vj but not to vi.

We consider G′ where
V(G′) = V(G0) and E(G′) = (E(G0)∖{vivn, vjvk}) ∪ {vjvn, vivk}.

Then G′ has score D too, and j(G′) ≤ j(G0) − 1, a contradiction.


This proves the claim.






You might also like