1

Data Structures and Algorithms

BITS-Pilani K. K. Birla Goa Campus

1
Material for the presentation taken from Cormen, Leiserson, Rivest and
Stein, Introduction to Algorithms, Third Edition; 1
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Sorting algorithms

I In place sorting : If at any time only a constant number of

elements are stored outside the array.

2
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Sorting algorithms

I In place sorting : If at any time only a constant number of

elements are stored outside the array.
Auxilliary space complexity is

2
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Sorting algorithms

I In place sorting : If at any time only a constant number of

elements are stored outside the array.
Auxilliary space complexity is Θ(1)

2
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Sorting algorithms

I In place sorting : If at any time only a constant number of

elements are stored outside the array.
Auxilliary space complexity is Θ(1)
I MERGE procedure does not operate in place.

2
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Sorting algorithms

I In place sorting : If at any time only a constant number of

elements are stored outside the array.
Auxilliary space complexity is Θ(1)
I MERGE procedure does not operate in place.
I Ch 6 : Heapsort that uses a data structure called heap.

2
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Sorting algorithms

I In place sorting : If at any time only a constant number of

elements are stored outside the array.
Auxilliary space complexity is Θ(1)
I MERGE procedure does not operate in place.
I Ch 6 : Heapsort that uses a data structure called heap.
I Heapsort sorts n elements in place in O(n lg n) time.

2
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Binary tree

I Binary tree is a tree data structure where each node can have
at most two child nodes : left child node and right child node.

3
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Full Binary tree

I Each node is either a leaf node or has two child nodes.

4
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Complete Binary tree

I A complete binary tree is a binary tree in which all levels are

completely filled except possibly the last level, which is filled
from left to right.

5
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Complete Binary tree

I A complete binary tree is a binary tree in which all levels are

completely filled except possibly the last level, which is filled
from left to right.

5
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Heap data structure : a complete binary tree

6
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Heap data structure : a complete binary tree

I The last level is not completely filled.

6
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Heap : a complete binary tree

I Each node of the heap corresponds to an element of the array.

7
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Heap : a complete binary tree

I Each node of the heap corresponds to an element of the array.

7
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Heap : a complete binary tree

I Each node of the heap corresponds to an element of the array.

I A.length,

7
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Heap : a complete binary tree

I Each node of the heap corresponds to an element of the array.

I A.length, A.heap-size,

7
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Heap : a complete binary tree

I Each node of the heap corresponds to an element of the array.

I A.length, A.heap-size, Root : A[1]

7
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Heap : parent, left child, right child

8
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Binary Heaps

I Either max-heaps or min-heaps

9
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Binary Heaps

I Either max-heaps or min-heaps

I Max-heap property: A[PARENT (i)] ≥ A[i]

9
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Binary Heaps

I Either max-heaps or min-heaps

I Max-heap property: A[PARENT (i)] ≥ A[i]

9
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Binary Heaps

I Either max-heaps or min-heaps

I Max-heap property: A[PARENT (i)] ≥ A[i]

I A[1] contains the maximum element

9
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Binary Heaps

I Min heap

10
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Binary Heaps

I Min heap
I Min-heap property: A[PARENT (i)] ≤ A[i]

10
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Binary Heaps

I Min heap
I Min-heap property: A[PARENT (i)] ≤ A[i]
I A[1] will contain the smallest element

10
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Heap

I Height of a node : number of edges on the longest simple

downward path from the node to a leaf.

11
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Heap

I Height of a node : number of edges on the longest simple

downward path from the node to a leaf.

11
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Heap

I Height of a node : number of edges on the longest simple

downward path from the node to a leaf.

I Height of a heap is the height of its root.

11
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Height of a heap

I Suppose a heap has n elements and has a height of h.

12
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Height of a heap

I Suppose a heap has n elements and has a height of h.

12
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Height of a heap

I Suppose a heap has n elements and has a height of h.

2h

12
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Height of a heap

I Suppose a heap has n elements and has a height of h.

2h ≤ n

12
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Height of a heap

I Suppose a heap has n elements and has a height of h.

2h ≤ n ≤ 2h+1 − 1

12
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Height of a heap

I Suppose a heap has n elements and has a height of h.

2h ≤ n ≤ 2h+1 − 1
2h ≤ n < 2h+1

12
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Height of a heap

I Suppose a heap has n elements and has a height of h.

2h ≤ n ≤ 2h+1 − 1
2h ≤ n < 2h+1
h ≤ lg n < (h + 1)

12
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Height of a heap

I Suppose a heap has n elements and has a height of h.

2h ≤ n ≤ 2h+1 − 1
2h ≤ n < 2h+1
h ≤ lg n < (h + 1)
h = blg nc
12
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
MAX-HEAPIFY: Maintaining the heap property

I MAX-HEAPIFY(A,i) : binary trees rooted at LEFT(i) and

RIGHT(i) satisfy max-heap property.

13
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
MAX-HEAPIFY: Maintaining the heap property

I MAX-HEAPIFY(A,i) : binary trees rooted at LEFT(i) and

RIGHT(i) satisfy max-heap property.
https://visualgo.net/en/heap

13
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
MAX-HEAPIFY: Maintaining the heap property

14
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
MAX-HEAPIFY: Maintaining the heap property

15
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
MAX-HEAPIFY: Finding recurrence

I Let the tree rooted at i have n nodes. The child subtree will
have a size at most 2n/3

16
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
MAX-HEAPIFY: Finding recurrence

I Let the tree rooted at i have n nodes. The child subtree will
have a size at most 2n/3

16
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
MAX-HEAPIFY: Finding recurrence

I Let the tree rooted at i have n nodes. The child subtree will
have a size at most 2n/3

h = height of node i
Maximum size of child subtree = 2h − 1

16
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
MAX-HEAPIFY: Finding recurrence

I Let the tree rooted at i have n nodes. The child subtree will
have a size at most 2n/3

h = height of node i
Maximum size of child subtree = 2h − 1

Maximum fraction of nodes =

16
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
MAX-HEAPIFY: Finding recurrence

I Let the tree rooted at i have n nodes. The child subtree will
have a size at most 2n/3

h = height of node i
Maximum size of child subtree = 2h − 1
2h − 1
Maximum fraction of nodes =
2h − 1 + 2h−1
16
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
MAX-HEAPIFY: Finding recurrence

2h − 1
Maximum fraction of nodes =
2h − 1 + 2h−1

17
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
MAX-HEAPIFY: Finding recurrence

2h − 1
Maximum fraction of nodes =
2h − 1 + 2h−1
A useful fractional inequality:
a a+1
For 0 < a < b , <
b b+1

17
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
MAX-HEAPIFY: Finding recurrence

2h − 1
Maximum fraction of nodes =
2h − 1 + 2h−1
A useful fractional inequality:
a a+1
For 0 < a < b , <
b b+1
h
2 −1
<
2h − 1 + 2h−1

17
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
MAX-HEAPIFY: Finding recurrence

2h − 1
Maximum fraction of nodes =
2h − 1 + 2h−1
A useful fractional inequality:
a a+1
For 0 < a < b , <
b b+1
h
2 −1 2h
<
2h − 1 + 2h−1 2h + 2h−1

17
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
MAX-HEAPIFY: Finding recurrence

2h − 1
Maximum fraction of nodes =
2h − 1 + 2h−1
A useful fractional inequality:
a a+1
For 0 < a < b , <
b b+1
h
2 −1 2h
<
2h − 1 + 2h−1 2h + 2h−1
2 × 2h−1
=
(2 + 1) × 2h−1

17
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
MAX-HEAPIFY: Finding recurrence

2h − 1
Maximum fraction of nodes =
2h − 1 + 2h−1
A useful fractional inequality:
a a+1
For 0 < a < b , <
b b+1
h
2 −1 2h
<
2h − 1 + 2h−1 2h + 2h−1
2 × 2h−1 2
= h−1
=
(2 + 1) × 2 3

17
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
MAX-HEAPIFY: Finding recurrence

2h − 1
Maximum fraction of nodes =
2h − 1 + 2h−1
A useful fractional inequality:
a a+1
For 0 < a < b , <
b b+1
h
2 −1 2h
<
2h − 1 + 2h−1 2h + 2h−1
2 × 2h−1 2
= h−1
=
(2 + 1) × 2 3
2n
Maximum size of child subtree <
3

17
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
MAX-HEAPIFY: Running time

I T (n) ≤ T (2n/3) + c

18
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
MAX-HEAPIFY: Running time

I T (n) ≤ T (2n/3) + c
log 3 1
nlogb a = n 2 =1

18
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
MAX-HEAPIFY: Running time

I T (n) ≤ T (2n/3) + c
log 3 1
nlogb a = n 2 =1
logb a
c = Θ(n lgk n)

18
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
MAX-HEAPIFY: Running time

I T (n) ≤ T (2n/3) + c
log 3 1
nlogb a = n 2 =1
logb a
c = Θ(n lgk n)
= Θ(1) , for k = 0

18
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
MAX-HEAPIFY: Running time

I T (n) ≤ T (2n/3) + c
log 3 1
nlogb a = n 2 =1
logb a
c = Θ(n lgk n)
= Θ(1) , for k = 0
Soln. Θ(nlogb a lgk+1 n)

18
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
MAX-HEAPIFY: Running time

I T (n) ≤ T (2n/3) + c
log 3 1
nlogb a = n 2 =1
logb a
c = Θ(n lgk n)
= Θ(1) , for k = 0
Soln. Θ(nlogb a lgk+1 n) = Θ(lg n)

18
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
MAX-HEAPIFY: Running time

I T (n) ≤ T (2n/3) + c
log 3 1
nlogb a = n 2 =1
logb a
c = Θ(n lgk n)
= Θ(1) , for k = 0
Soln. Θ(nlogb a lgk+1 n) = Θ(lg n)
T (n) = O(lg n)

18
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
BUILD-MAX-HEAP : Building a heap

I Convert an unordered array into a max-heap using

MAX-HEAPIFY

19
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
BUILD-MAX-HEAP : Building a heap

I Convert an unordered array into a max-heap using

MAX-HEAPIFY

19
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
BUILD-MAX-HEAP : Building a heap

I Convert an unordered array into a max-heap using

MAX-HEAPIFY

What will be the parent node of the last element in array A?

19
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
BUILD-MAX-HEAP : Building a heap

I Convert an unordered array into a max-heap using

MAX-HEAPIFY

What will be the parent node of the last element in array A?

bA.length/2c

19
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
BUILD-MAX-HEAP : Building a heap

20
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
BUILD-MAX-HEAP : Building a heap

20
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
BUILD-MAX-HEAP : Building a heap

20
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
BUILD-MAX-HEAP : Running time

21
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
BUILD-MAX-HEAP : Running time

Running time:

21
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
BUILD-MAX-HEAP : Running time

Running time: O(n lg n)

21
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
BUILD-MAX-HEAP : Running time

Running time: O(n lg n) { Not asymptotically tight }

21
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
 l n m
Nodes with height h ≤ h+1
2

22
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
 l n m
Nodes with height h ≤ h+1
2

l n m
I There are at most nodes having a height h.
2h+1

22
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
 l n m
Nodes with height h ≤ h+1
2

l n m
I There are at most nodes having a height h.
2h+1

h = 0,

22
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
 l n m
Nodes with height h ≤ h+1
2

l n m
I There are at most nodes having a height h.
2h+1
 
10
h = 0,
20+1

22
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
 l n m
Nodes with height h ≤ h+1
2

l n m
I There are at most nodes having a height h.
2h+1
 
10
h = 0, = 5;
20+1

22
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
 l n m
Nodes with height h ≤ h+1
2

l n m
I There are at most nodes having a height h.
2h+1
 
10
h = 0, = 5; h = 1,
20+1

22
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
 l n m
Nodes with height h ≤ h+1
2

l n m
I There are at most nodes having a height h.
2h+1
   
10 10
h = 0, = 5; h = 1, =3
20+1 21+1

22
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
 l n m
Nodes with height h ≤ h+1
2

l n m
I There are at most nodes having a height h.
2h+1
   
10 10
h = 0, = 5; h = 1, =3
20+1 21+1

h = 2,

22
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
 l n m
Nodes with height h ≤ h+1
2

l n m
I There are at most nodes having a height h.
2h+1
   
10 10
h = 0, = 5; h = 1, =3
20+1 21+1
 
10
h = 2, = 2;
22+1
22
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
 l n m
Nodes with height h ≤ h+1
2

l n m
I There are at most nodes having a height h.
2h+1
   
10 10
h = 0, = 5; h = 1, =3
20+1 21+1
 
10
h = 2, = 2; h = 3,
22+1
22
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
 l n m
Nodes with height h ≤ h+1
2

l n m
I There are at most nodes having a height h.
2h+1
   
10 10
h = 0, = 5; h = 1, =3
20+1 21+1
   
10 10
h = 2, = 2; h = 3, =1
22+1 23+1
22
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
 l n m
Nodes with height h ≤ h+1
2
I Proof by Induction

23
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
 l n m
Nodes with height h ≤ h+1
2
I Proof by Induction
I Base case: h = 0

23
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
 l n m
Nodes with height h ≤ h+1
2
I Proof by Induction
I Base case: h = 0 lnm
Number of nodes having zero height =
2

23
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
 l n m
Nodes with height h ≤ h+1
2
I Proof by Induction
I Base case: h = 0 lnm l n m
Number of nodes having zero height = ≤ h+1
2 2

23
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
 l n m
Nodes with height h ≤ h+1
2
I Proof by Induction
I Base case: h = 0 lnm l n m
Number of nodes having zero height = ≤ h+1
2 2
I Inductive step: l n m
Assume that there are at most (h−1)+1 number of nodes
2
of height h − 1.

23
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
 l n m
Nodes with height h ≤ h+1
2
I Proof by Induction
I Base case: h = 0 lnm l n m
Number of nodes having zero height = ≤ h+1
2 2
I Inductive step: l n m
Assume that there are at most (h−1)+1 number of nodes
2
of height h − 1.
Let kh be the number of nodes of height h in the binary heap
T.

23
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
 l n m
Nodes with height h ≤ h+1
2
I Proof by Induction
I Base case: h = 0 lnm l n m
Number of nodes having zero height = ≤ h+1
2 2
I Inductive step: l n m
Assume that there are at most (h−1)+1 number of nodes
2
of height h − 1.
Let kh be the number of nodes of height h in the binary heap
T.
Let us construct a new heap T 0 by removing all leaf nodes
from T

23
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
 l n m
Nodes with height h ≤ h+1
2
I Proof by Induction
I Base case: h = 0 lnm l n m
Number of nodes having zero height = ≤ h+1
2 2
I Inductive step: l n m
Assume that there are at most (h−1)+1 number of nodes
2
of height h − 1.
Let kh be the number of nodes of height h in the binary heap
T.
Let us construct a new heap T 0 by removing all leaf nodes
from T
0
kh = kh−1

23
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
 l n m
Nodes with height h ≤ h+1
2
I Proof by Induction
I Base case: h = 0 lnm l n m
Number of nodes having zero height = ≤ h+1
2 2
I Inductive step: l n m
Assume that there are at most (h−1)+1 number of nodes
2
of height h − 1.
Let kh be the number of nodes of height h in the binary heap
T.
Let us construct a new heap T 0 by removing all leaf nodes
from T
n0
 
0
kh = kh−1 ≤ (h−1)+1
2

23
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
 l n m
Nodes with height h ≤ h+1
2
I Proof by Induction
I Base case: h = 0 lnm l n m
Number of nodes having zero height = ≤ h+1
2 2
I Inductive step: l n m
Assume that there are at most (h−1)+1 number of nodes
2
of height h − 1.
Let kh be the number of nodes of height h in the binary heap
T.
Let us construct a new heap T 0 by removing all leaf nodes
from T
n0
   
0 b(n/2)c
kh = kh−1 ≤ (h−1)+1 = (h−1)+1
2 2

23
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
 l n m
Nodes with height h ≤ h+1
2
I Proof by Induction
I Base case: h = 0 lnm l n m
Number of nodes having zero height = ≤ h+1
2 2
I Inductive step: l n m
Assume that there are at most (h−1)+1 number of nodes
2
of height h − 1.
Let kh be the number of nodes of height h in the binary heap
T.
Let us construct a new heap T 0 by removing all leaf nodes
from T
n0
   
0 b(n/2)c
kh = kh−1 ≤ (h−1)+1 = (h−1)+1
2 2
 
(n/2)
≤
2h
23
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
 l n m
Nodes with height h ≤ h+1
2
I Proof by Induction
I Base case: h = 0 lnm l n m
Number of nodes having zero height = ≤ h+1
2 2
I Inductive step: l n m
Assume that there are at most (h−1)+1 number of nodes
2
of height h − 1.
Let kh be the number of nodes of height h in the binary heap
T.
Let us construct a new heap T 0 by removing all leaf nodes
from T
n0
   
0 b(n/2)c
kh = kh−1 ≤ (h−1)+1 = (h−1)+1
2 2
  l
(n/2) n m
≤ =
2h 2h+1
23
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
BUILD-MAX-HEAP : Running time

24
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
BUILD-MAX-HEAP : Running time

blg nc l
X n m
h+1
O(h)
2
h=0

24
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
BUILD-MAX-HEAP : Running time

 
blg nc l blg nc
X n m X h
O(h) = O n
2h+1 2h
h=0 h=0

24
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
BUILD-MAX-HEAP : Running time

 
blg nc l blg nc
X n m X h
O(h) = O n
2h+1 2h
h=0 h=0
∞
X
kx k
k=0

24
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
BUILD-MAX-HEAP : Running time

 
blg nc l blg nc
X n m X h
O(h) = O n
2h+1 2h
h=0 h=0
∞
X x
kx k =
(1 − x)2
k=0

24
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
BUILD-MAX-HEAP : Running time

 
blg nc l blg nc
X n m X h
O(h) = O n
2h+1 2h
h=0 h=0
∞
X x
kx k =
(1 − x)2
k=0
∞
X
k(1/2)k
k=0
24
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
BUILD-MAX-HEAP : Running time

 
blg nc l blg nc
X n m X h
O(h) = O n
2h+1 2h
h=0 h=0
∞
X x
kx k =
(1 − x)2
k=0
∞
X 1/2
k(1/2)k =
(1 − (1/2))2
k=0
24
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
BUILD-MAX-HEAP : Running time

 
blg nc l blg nc
X n m X h
O(h) = O n
2h+1 2h
h=0 h=0
∞
X x
kx k =
(1 − x)2
k=0
∞
X 1/2
k(1/2)k = =2
(1 − (1/2))2
k=0
24
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
BUILD-MAX-HEAP : Running time

 
blg nc l blg nc
X n m X h
O(h) = O n 
2h+1 2h
h=0 h=0
∞
!
X h
=O n
2h
h=0

24
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
BUILD-MAX-HEAP : Running time

 
blg nc l blg nc
X n m X h
O(h) = O n 
2h+1 2h
h=0 h=0
∞
!
X h
=O n
2h
h=0
= O(2n)

24
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
BUILD-MAX-HEAP : Running time

 
blg nc l blg nc
X n m X h
O(h) = O n 
2h+1 2h
h=0 h=0
∞
!
X h
=O n
2h
h=0
= O(2n) = O(n)

24
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
P∞ k x
k=0 kx = (1−x)2

25
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
P∞ k x
k=0 kx = (1−x)2

I 0<x <1, 1 + x + x2 + . . .

25
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
P∞ k x
k=0 kx = (1−x)2

1
I 0<x <1, 1 + x + x2 + . . . =
1−x

25
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
P∞ k x
k=0 kx = (1−x)2

1
I 0<x <1, 1 + x + x2 + . . . =
1−x
∞
X 1
I xk =
1−x
k=0

25
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
P∞ k x
k=0 kx = (1−x)2

1
I 0<x <1, 1 + x + x2 + . . . =
1−x
∞
X 1
I xk =
1−x
k=0
I Differentiating both sides w.r.t x
∞
X 1
kx k−1 =
(1 − x)2
k=0

25
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
P∞ k x
k=0 kx = (1−x)2

1
I 0<x <1, 1 + x + x2 + . . . =
1−x
∞
X 1
I xk =
1−x
k=0
I Differentiating both sides w.r.t x
∞
X 1
kx k−1 =
(1 − x)2
k=0

I Multiplying x on both sides

∞
X x
kx k =
(1 − x)2
k=0

25
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Heap

I We can construct a max heap (or a min heap) from an

unordered array in O(n) time.

26
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Heap

I We can construct a max heap (or a min heap) from an

unordered array in O(n) time.
I What is the running time of Build-Max-Heap procedure in
Θ-notation?

26
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Heapsort

27
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Heapsort

I Running time : O(n lg n)

27
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Heapsort

I Running time : O(n lg n)

I Operation (P. 161)

27
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Priority Queues

I Problem: Data structure for a set of Keys such that the

following operations can be performed in O(lg n) time:

28
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Priority Queues

I Problem: Data structure for a set of Keys such that the

following operations can be performed in O(lg n) time:
1. Maximum(S)

28
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Priority Queues

I Problem: Data structure for a set of Keys such that the

following operations can be performed in O(lg n) time:
1. Maximum(S)
2. Extract-Max(S)

28
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Priority Queues

I Problem: Data structure for a set of Keys such that the

following operations can be performed in O(lg n) time:
1. Maximum(S)
2. Extract-Max(S)
3. Increase-Key(S,x,k)

28
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Priority Queues

I Problem: Data structure for a set of Keys such that the

following operations can be performed in O(lg n) time:
1. Maximum(S)
2. Extract-Max(S)
3. Increase-Key(S,x,k)
4. Insert(S,x)

28
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Priority Queues

I Problem: Data structure for a set of Keys such that the

following operations can be performed in O(lg n) time:
1. Maximum(S)
2. Extract-Max(S)
3. Increase-Key(S,x,k)
4. Insert(S,x)
I We will use the Heap data structure to construct a
max-priority queue.

28
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Priority Queues

I Problem: Data structure for a set of Keys such that the

following operations can be performed in O(lg n) time:
1. Maximum(S)
2. Extract-Max(S)
3. Increase-Key(S,x,k)
4. Insert(S,x)
I We will use the Heap data structure to construct a
max-priority queue.
I Used in Graph algorithms (e.g. Dijkstra’s algorithm)

28
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Priority Queues

29
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Priority Queues

29
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Priority Queues

Running time of Heap-Extract-Max : O(lg n)

29
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Max-Heap-Increase-Key

30
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Max-Heap-Increase-Key

I Suppose we increase the value at index i to 15.

30
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Max-Heap-Increase-Key

I Suppose we increase the value at index i to 15.

How can we fix the Max-Heap property?

30
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Max-Heap-Increase-Key

31
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Max-Heap-Increase-Key

I Running time of Max-Heap-Increase-Key :

31
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Max-Heap-Increase-Key

I Running time of Max-Heap-Increase-Key : O(lg n)

31
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Max-Heap-Insert

32
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Max-Heap-Insert

Running time of Max-Heap-Insert :

32
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Max-Heap-Insert

Running time of Max-Heap-Insert : O(lg n)

32
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Priority Queues

I Using a max-heap, all the basic operations can be performed

in O(lg n) time.

33
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Priority Queues

I Using a max-heap, all the basic operations can be performed

in O(lg n) time.
I Two forms : max-priority queue and min priority-queue

33
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Priority Queues

I Using a max-heap, all the basic operations can be performed

in O(lg n) time.
I Two forms : max-priority queue and min priority-queue
I Min-priority queue can be implemented using a min-heap.

33
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms
Priority Queues

I Using a max-heap, all the basic operations can be performed

in O(lg n) time.
I Two forms : max-priority queue and min priority-queue
I Min-priority queue can be implemented using a min-heap.
I Min-priority queue operations:
1. Minimum(S)
2. Extract-Min(S)
3. Decrease-Key(S,x,k)
4. Insert(S,x)

33
BITS-Pilani K. K. Birla Goa Campus Data Structures and Algorithms