Graphs

Data Structures and Algorithms

Outline
 Non-Linear Data Structures
 Graphs
 Intro
 Application
 Terminologies
 Representation
 Adjacency List
 Adjacency Matrix

2
Graph
 A graph is a way of representing relationships between pairs of objects.
 It is a set of objects, called vertices, together with a collection of pairwise
connections between them, called edges

3
Graph
 Graph is a mathematical structure that is defined as G= (v, e), where v is a set
of vertices{v1, v2, …vn} and e is a set of edges {e1,e2,e3,…em}
 Where edge e is an ordered pair of two vertices, represents a connection between two vertices
 Graph can be directed, or undirected
 Example:
 V = {A, B, C, D, E, F, G}
 E = { {A, B}, {A, D}, {A, E}, {B, C}, {B, D}, {B, E}, {C, E}, {C, F}, {D, E} }
 Then the graph G=(V,E) is: B B
 |V|= 7 A A
 C C
|E|=9

D E F G D F G
E
4
Applications
 Maps

 Communication Networks Social networks

5
Applications
 Flight routes

 Web graphs Maze (games)

6
Terminologies
 Directed Graph  Un-Directed Graph
 Also called digraph  Edge has no direction and considered
 Every edge has a direction, an incoming as two-ways
edge is not equal to outgoing edge  Vertex order is not important
V

U X Z

 Vertex order is important

7
Terminologies
 End Points  Parallel Edges
 Vertices at both ends of an edge
 Edges with same end points
 Incident Edge
 h and i are parallel edges
 If vertex is an end point of edge, edge is incident on
that vertex
 a is incident on u and v
 Adjacent/Neighbor Vertices V
a b
 Vertices that are end points of same edge h
 u, v and w are adjacent d
 Self Loop U X Z j
 Node connected to itself c e i
 Z is in self loop and j is self edge
 S W g
Degree of Node
 Number of incident edges f
 U has degree 2, X has degree 5

Y
Self edge is considered twice, so z has degree 4
8
Terminologies
 In a directed graph we can distinguish between
 Incoming Edges
 Directed edges for which the given vertex is destination
 c is incoming edge of U
 Outgoing Edges
 Directed edges for which the given vertex is origin
 a is outgoing edge of U V
a b
 In-Degree of Vertex h
 Number of incoming edges d
 Out-Degree of Vertex U X Z
 Number of outgoing edges c e
 Source Vertices i
 Vertices with an in-degree of zero W g
 w is source vertex
 Sink Vertices f
 Vertices with an out-degree of zero Y
 v is sink vertex
9
Terminologies
 Path
 A path between two vertices is sequence of alternating vertices and edges, where each
successive vertex is connected.

V
V

U X Z
U X Z

W W

Y Y

10
Terminologies
 Simple Path
 A path with distinct nodes
 One of the possible paths from u to v is u-w-v
 Cycle
 A path that starts and finish at same vertex V
V

U X Z
U X Z

W W

Y Y

11
Terminologies
 Reachability of Vertex
 A vertex v is reachable from other vertex if there exists a path from other vertex to
vertex v
 In undirected graph, edge is considered as 2-ways, so every vertex is reachable in following
example.

V V

U X Z U X Z

 U , V and X are not reachable from Z all vertices are reachable from each other

12
Terminologies
 Connected Graph  Disconnected Graph
 Each vertex is reachable from every other  A graph that is not connected
vertex, in other words there exists a path
from each vertex to every other vertex

V V

U X Z U X Z

13
Terminologies
 Strongly Connected Graph  Weakly Connected Graph
 If there exists a directed path between  If there exists a path between each pair
each pair of vertices of vertices which ignores direction

V V

U X Z U X Z

14
Terminologies
 Completely Connected Graph
 If there exists an edge between each pair of vertices
 Directed  Undirected
V

V
U X

U X
X

 Maximum Edges?  Maximum Edges

 |E|=|V|*|V-1| =O(|V|2)  |E|=|V|*|V-1|/2= O(|V|2)
15
Terminologies
 Sub Graph V
 A graph that is consists of subset of vertices and edges of G

 U X Z
Two possible sub-graphs
V

U X Z
U X

 Following is not valid sub-graph of G, why?

 Look at definition of sub-graph

U X Z

16
Terminologies
 Simple Graph
 A graph with no parallel and self edges
V

U X Z

17
Terminologies
 Weighted Graph  Non-Weighted Graph
 Weights on edge means cost or distance
between end points of edge

18
Terminologies
 Path Length
 Sum of weights of edges on path from one vertex to other.
 Length of path between u and w is 2
 Length of path between u and y is 3

V
There are two paths from u to x 1 2
2
 Path u-w-x has length 6 and path u-w-y-x has length 5
2
 Shortest Path U X Z
 Path with minimum length 2 4 1
 From u to x is 5
W 2
 From u to v is 4
1
Y

19
Terminologies
 Tree
 An undirected graph G is tree if it fulfills any of the following condition:
 G has V-1 edges and no cycles
 G has V-1 edges and is connected
 G is connected, but removing any edge disconnects it
 G is acyclic, but adding any edges creates a cycle
 Exactly one simple path between each pair of vertices in G

V V

U X Z U X Z

20
Terminologies
 Directed Acyclic Graph (DAG)
 Directed graph without cycles
V

U X Z

 Following graph is not DAG, as it contains a cycle

U X Z

21
Terminologies
 Directed Acyclic Graph (DAG) B D
 A directed graph without cycles
A
 Applications: C E
 The parse tree constructed by a compiler to execute sequential statements
 Dependency graphs for task scheduling
 Dependency graphs between classes formed by inheritance relationships in object-oriented
programming languages
 Information categorization systems, such as folders in a computer
 Course pre-requisites CS16 CS141 CS242
CS15

CS17 CS32
CS18 CS123 CS224

22
Graph ADT
 Common methods for Graph ADT can be:
 numVertices()
 vertices()
 numEdges()
 edges()
 outgoingEdges(v)
 incomingEdges(v)
 getEdge(v1, v2)
 endVertices(e)
 opposite(v, e)
 insertVertex(value)
 insertEdge(v1, v2, value)
 removeVertex(v)
 removeEdge(e)

 Run time depends upon underlying implementation

23
Memory Representation
 There are multiple ways to represent a graph in memory:
 Adjacency Matrix
 Adjacency List
 Edge List
 Adjacency Map

 Assume following example for all representations

V1 V3

V5

V2 V4

24
Adjacency Matrix
 The adjacency matrix of a graph G = (V, E) is a |V | × |V | matrix E, where each
entry Eij is equal to 1 if there exists an edge e = (vi, vj ) ∈ E and 0 otherwise.
 1 and 0 can also be replaced with true/false.
 Vertex list itself is stored in 1D array
0 1 2 3 4
“V1” “V2” “V3” “V4” “V5”

V1 V2 V3 V4 V5
V1 0 1 1 0 0
V1 V3 V2 1 0 1 1 0

V5
V3 1 1 0 1 1
V4 0 1 1 0 1
V2 V4 V5 0 0 1 1 0

25
Adjacency Matrix
 If graph is directed?
 Then entry Eij is equal to 1 if there exists an edge e = from vi to vj and 0 otherwise.
V1 V2 V3 V4 V5
V1 0 1 1 0 0
V1 V3
V2 0 0 1 0 0
V3 0 0 0 1 1 V5

V4 0 1 0 0 0
V2 V4
V5 0 0 0 1 0

 Weighted Graph:
 1 and 0 are replaced with respective weights
 0 or -1 presents no edge

26
Example

27
Example
 If vertices labels can be used as index, then separate 1D array is not needed to
store them, edge matrix would be enough.
0 1 2 3 4
0 0 1 1 0 0
0 2
1 0 0 1 0 0
2 0 0 0 1 1 4
3 0 1 0 0 0
1 3
4 0 0 0 1 0

28
Adjacency Matrix
 Vertex list and Edge list are given
 0 1 2 3 4
Running Time?
 “V1” “V2” “V3” “V4” “V5”
Get a vertex’s out-edges
 Get a vertex’s in-edges V1 V2 V3 V4 V5
 Decide if some edge exists
V1 0 1 1 0 0
 Insert an edge
V2 0 0 1 0 0
 Delete an edge

V3 0 0 0 1 1
Inset a vertex
 V4 0 1 0 0 0
Remove a vertex
 V5 0 0 0 1 0
Memory
 O(|V|2)
 Good for?
 Dense graphs
 Edges are close to |V|2
29
Adjacency List
 Each vertex is associated with its list of connected vertices.
 There can be different ways to implement it. Simplest one is using vertex numbers as
array index, where each index would hold a pointer to linked list of connected vertices.

0 1 2
0 2
1 0 2 3
4
2 0 2 3 4
1 3
3 1 2 4

4 2 3

30
Adjacency List
 Another idea can be to store edges list inside of vertex and storing vertices in array.
 Edge list then can be a list of vertices itself or list of indices of connected vertices
 It also can be label of vertices

1 “V1” 2 3

2 “V2” 1 3 4 V1 V3

3 “V3” 1 2 4 5 V5

4 “V4” 2 3 5 V2 V4

5 “V5” 3 4

 Here edge list is storing indicies

31
32
Adjacency List
 Running Time?
 1 “V1” 2 3
Get a vertex’s out-edges
 Get a vertex’s in-edges 2 “V2” 1 3 4
 Decide if some edge exists
 3 “V3” 1 2 4 5
Insert an edge
 Delete an edge 4 “V4” 2 3 5
 Inset a vertex
 5 “V5” 3 4
Remove a vertex
 Memory
 O(|V|+|E|)
 Good for?
 Sparse
 Edges are significantly less than |V| 2
33