Algorithms

Lecture 19
In the Last Lecture
O Greedy Algorithms
O Huffman Coding
 Graph
Algorithms
Graph Algorithm
O Definition
O Some Properties
O Graph Representation
O Graph Traversal
O BFS
O DFS
 Introduction
O Any time there is a set of objects
and there is some sort of
“connection” or relationship or
interaction between pairs of
objects
O Graph is a good way to model
this
O Computer and communication
networks
O Transportation networks
 Introduction
O VLSI, logic circuits
O Surface meshes for shape
description in computer aided
design and GIS
O Precedence constraints
O A graph G=(V,E) consists of:
O A finite set of vertices v(or nodes)
O E, a binary relation u called edges
O If set of pairs are ordered we have
directed graph or digraph
undirected graph digraph

1 2
1 2

3 4 3 4

1
G=(V,E)
V={1,2,3,4}
3 4
E={(1,2),(2,1),(2,4),(4,2)..}
Self loop
 Adjacent Vertices
O A vertex w is adjacent to vertex v if
there is an edge from v to w
O The number of edges coming out of a
vertex is called the out-degree of that
vertex

1 2 1 and 2
1 and 3
1 and 4
3 4 3 and 4
 Graph Properties
O The number of edges coming in is
the in-degree of that vertex
O A path in a un-direct graph is a
sequence of vertices and length of
the path is number of edges
O A vertex w is reachable from
vertex u if there is a path from v
to w
O A cycle in a digraph is a path
containing at least one edge
 Graph Properties
O Hamiltonian cycle is a cycle that visits
every vertex of in a graph exactly once
O An Eulerian cycle is a cycle that visits
every edge of graph exactly once

1-2-4-3-1
1 2
1-4-3-1
1-3-1

3 4
 Graph Properties
O A graph is said to be acyclic if it
contains no cycles
O A graph is connected if every vertex
can reach every other vertex
O A directed graph that is acyclic is
called a direct acyclic graph(DAG)
O A weighted graph is a triple(V,E,W)
W(e) is called the weight of e
 Graph Representation
O We use two methods to
represent graphs on computer
i. Adjacency-list
representation
ii. Adjacency-matrix
representation
O Adjacency-list consist of an
array of |V| lists, one for each
vertex
O Adjacency-list of a graph
requires Θ(V+E) of memory
Adjacency-list representation
1 2

5 4

1 2 5 /

2 1 3 4 /
3 2 4 /
4 2 5 3 /

5 4 1
Adjacency-list representation
1 2 3

4 5 6

1 2 4 /

2 5 /
3 6 5 /
4 2 /

5 4 /

6 6 /
Adjacency-matrix representation

OAdjacency-matrix is a |V|*|V|
OA=(aij) such that:
1 if(i,j)ε E
aij=
0 otherwise

OAdjacency-matrix of a graph
requires Θ(V2) of memory
 Adjacency-matrix representation

1 2 3 4 5
1 2
1 0 1 0 0 1
3
2 1 0 1 1 1
5 4
3 0 1 0 1 0
4 0 1 1 0 1
5 1 1 0 1 0
Adjacency-matrix representation
1 2 3 4 5 6

1 2 3
1 0 1 0 1 0 0
2 0 0 0 0 1 0
4 5 6 3 0 0 0 0 1 1
4 0 1 0 0 0 0
5 0 0 0 1 0 0
6 0 0 0 0 0 1
 Which is better?
O It depends on the type of
application
O Adjacency-list saves the
memory
O But it is difficult to determine
whether there is a path
between two nodes or not
O This aspect can be easily
covered in adjacency-matrix
but it suffers as far as space is
concerned
 Graph Traversal
O In general, there is no natural “first”
node in graph as root in tree
O There is no natural order among
successors of a particular node
O Unlike, tree a node in graph may have
more than one predecessor
O Two traversal techniques are used
i. Breath first Search
ii. Depth first Search
 Graph Traversal
O For traversal we have to define a
source vertex s
O For each vertex vεV we will store the
d[v] which is the distance from s to v
note that d[s]=0
O We will also store a predecessor(or
parent) ∏[v] which is the first vertex
along this path if we walk from v
backward to s .We will also set
∏[s]=0
Breath First Search(BFS)

O Start with s and visit its adjacent

nodes
O Label the distance 1
O Now consider the neighbors of the
neighbors of s. They would be at
distance 2.
O Consider the neighbors of the
neighbors of the neighbors of s and
give them number 3
O Repeat this process until no more
unvisited neighbor left
 Breath First Search(BFS)
O BFS computes the distance(smallest
no. of edges from s to each vertex)
O It produces a breath first tree with
a root s that contain all reachable
vertices
O To keep track of the progress, BFS
colors each vertex white, gray and
then black
O Color of each vertex is stored
color[u]
O BFS(G,s)
1. for each vertex u ε V[G]-
{s}
2. do color[u]  WHITE
3. d[u] ∞
4. ∏[u] Nil
5. color[s]GRAY
6. d[s] 0
7. ∏[s] Nil
8. QΦ
9. Enqueue(Q,s)
10.While Q≠Φ
11. do uDequeue(Q)
12. for each v ε Adj[u]
13. do if color[v]=WHITE
14. then color[v] GRAY
15. d[v] d[u]+1
16. ∏[v]u
17. Enqueue(Q,v)
18. color[u]BLUE
 r s t u

(a)

v w x y

r s t u
∞ ∞ ∞
Q Φ
(b)
∞ ∞ ∞ ∞
v w x y
 r s t u
∞ 0 ∞ ∞
Q s Φ
(c)
0
∞ ∞ ∞ ∞
v w x y

r s t u
1 0 ∞ ∞

w r Φ Q
(d)
1 1
∞ 1 ∞ ∞
v w x y
 r s t u
1 0 2 ∞
Q r t x Φ
(e)
1 2 2
∞ 1 2 ∞
v w x y

r s t u
1 0 2 ∞

(f) Q t x v Φ

2 2 2
2 1 2 ∞
v w x y
 r s t u
1 0 2 3

(g) Q x v u Φ

2 1 2 ∞ 2 2 3
v w x y

r s t u
1 0 2 3

(h) Q v u y Φ

2 2 3
2 1 2 3
v w x y
 r s t u
1 0 2 3

(i) Q u y Φ

2 3
2 1 2 3
v w x y

r s t u
1 0 2 3

(j) Q y Φ

3
2 1 2 3
v w x y
 r s t u
1 0 2 3

Q Φ
(k)
2 1 2 3
v w x y
 Analysis
O BFS requires Θ (V+E) time to
complete its process
O Each item is only once enqueued
and dequeued
 Properties of BFS
O BFS creates a tree during it
traversal
O It calculates shortest path from
source to any other node if the
graph is un-weighted
 Depth-first Search
O The technique we use is called time-
stamp structure
O As we traverse the graph in DFS
order we will associate two numbers
with each vertex
O When we first visit a vertex v, store a
counter in d[v]
O When we finish processing a vertex
we store a counter in f[u]
O These two numbers are stamps
 d e

b a f

c g
 d e

b a f

1/..

c g
 d e

DFS-VISIT(a)
DFS-VISIT(b)
DFS-VISIT(c)
b a f

1/..
2/.. 1/..

c g

3/..
 d e

DFS-VISIT(a)
DFS-VISIT(b)
b a f DFS-VISIT(c)

2/5 1/..
2/5 1/..
c g

3/4
3/4
 d e

DFS-VISIT(a)
DFS-VISIT(f)
b a f
DFS-VISIT(g)
2/5 1/.. 6/..

c g

3/4 7/..
 d e

DFS-VISIT(a)

b a f DFS-VISIT(f)
DFS-VISIT(g)
2/5 1/.. 6/9

c g

3/4 7/8
 d e

b a f

2/5 1/10 6/9

c g

3/4 7/8
 d e

11/.. 12/..
DFS-VISIT(d)

b a f DFS-VISIT(e)

2/5 1/10 6/9

c g

3/4 7/8
 d e

11/.. 12/13
DFS-VISIT(d)

f DFS-VISIT(e)
b a

2/5 1/10 6/9

c g

3/4 7/8
 Properties of DFS
OEdge Classification
OParenthesis
Structure/Lemma
 Edge Classification
O For a directed graph, the non tree
edges can be classified as follows:
O Back edge:(u,v) where v is an
ancestor of u in the tree
O Forward edge:(u,v) where v is
proper descendent of u in the tree
O Cross edge:(u,v) where u and v
are disjoint.no ancestor and
decendent relationship
 d e

11/14 12/13

C C
b a f

2/5 1/10 6/9

F B
c g

3/4 7/8
c
 DFS Structure
O The time stamp given by the
DFS allow us to determine a
number of things about a graph
or digraph
O For example, we can determine
whether the graph contains any
cycle
O If there is any back edge in the
graph than it contains cycle
O Beware there is no relationship
between the no. of back edges
and no. of cycles
Parenthesis Structure
O If we represent (v as starting
time of vertex v and v) as
ending time
O DFS form a full parenthesized
structure
O Time of u is overlapping for v
OR
O Time of u and v are disjoint
 Applications of DFS
OTopological Sort
OStrongly Connected
Components
 Topological Sort
O A directed acyclic graph(DAG) arise
in many applications where there
are precedence or ordering
O There are series of tasks to be
performed and certain tasks must
precede other tasks
O In general, a precedence constraints
graph is a DAG in which vertices are
tasks and the edges(u,v) means
that task u must be completed
before task v begins
 Topological Sort
O A topological sort of a DAG is a
linear ordering of the vertices of
the DAG such that for each
edge(u,v) u appear before v in the
ordering
Computer Programming Discrete Math

Philosophy

Data structure D&A

Compiler Construction Theory of automata

Artificial Intelligence

Speech Recognition
 11/16 17/18

Computer Programming Discrete Math 9/10

Philosophy
12/15

Data structure D&A 13/14

6/7 1/8

Compiler Construction Theory of automata

2/5

Artificial Intelligence

3/4

Speech Recognition
 17/18 11/16 13/14
12/15
Discrete Math CP DS D&A

1/8 6/7
9/10
Philosophy Automata CC

2/5 3/4
Artificial Intelligence Speech Recognition
 Topological Sort
O TOPOLOGICAL-SORT(G)
1. Call DFS(G) to compute finishing
times f[v] for each vertex
2. As each vertex is finished ,insert
it onto the front of a linked list
3. Return the linked list of vertices
We can perform topological sort in
(V+E) time since DFS takes
(V+E) and it takes (1) time to
insert |V| vertices on to the front
of the link list
 Strongly Connected
Components
O Now, we consider an important
connectivity with digraph
O When digraphs are used in
communications and
transportations networks
people want to know there
networks are complete
O Complete in sense it is possible
to from any location to reach
any other location in a digraph
 Strong Connection
O A digraph is strongly connected if
for every pair of vertices u,vεV u
can reach v and vice-versa
O We will like to write an algorithm
that determines whether a digraph
is strongly connected or not
O In fact, we will solve a
generalization problem of this
problem, of computing the strongly
connected components of a digraph
 A

B C

D G I

E F H
 A

B C

D G I

E F H
 A,B,C

D,E

F,G,H,I

Component DAG
Algorithm
 Graph Transpose
O Graph transpose of a G=(V, E) is
O GT=(V,ET)

A A

C B C
B