⊖

SORTING
Algorithms

By
Kapil Yadav

Insertion Sort Merge Sort Selection Sort

Quick Sort Heap Sort Bubble Sort

 Bubble Sort:
Bubble Sort is a comparison based Sorting Algorithm.
It works by checking its adjacent element whether, it is in sorted order or not.
It is an Inplace Sorting algorithm as we don’t need any extra data structure while sorting.
It is stable (the sequence of repeating elements does not change).

Working of Bubble Sort:

15 16 6 8 5
A
7

[
15 16 6 8 5 6 8 5 15 16

§ &ˢ
"" " & 5 £ " " °
£96
" ""

15 6 16 8 5 6 5 8 15 16

IS 6 8 16 5 6 5 8 15 16

IS 6 8 5 16 6 5 8 15 16

IS 8 5 16 6 5 8 15 16

§
6

6 IS 8 5 16 5 6 8 15 16
Pass

, g ,, g ,, 5 6 8 15 16

6 8 5 15 16 5 6 8 15 16

6 8 5 15 16 5 6 8 15 16

No of Required Passes = (n-1)

Total Comparisons = (n-1) + (n-2) + (n-3) + ………….1 = ( (n-1) * n )/2 = O(n²) { Every Case}
Total Swaps :
minimum = 0 {elements are in sorted order}
maximum = (n-1) + (n-2) + (n-3) + ………….1 = ( (n-1) * n )/2 = O(n²)
Time Complexity = maximum(Comparison, Swaps)
= O(n²) {Every Case}
Space Complexity : O(1)

https://www.linkedin.com/in/kapilyadav22/
https://www.linkedin.com/in/kapilyadav22/
Further Optimization:
The Time Complexity of Bubble Sort can be optimized for Best Case.i.e When all the elements of array
are in sorted order, then there will be no swaps, So we can use it to identify whether our array is sorted
or not.

We can use a flag variable with initial value = 0,

If in any pass, there is no swap, just break the loop and we will get sorted elements.
TC: O(N) {Best Case}, O(n²) in {Average, Worst Case}

https://www.linkedin.com/in/kapilyadav22/
 https://www.linkedin.com/in/kapilyadav22/

Insertion Sort:

Insertion Sort is another Comparison Based Sorting Algorithm.

It works as similar as the way we sort playing cards.i.e we have unsorted and sorted portions of
the cards, and we use to place the unsorted card in its correct position.
Insertion Sort works similar to that.
It is also inplace Sorting algorithm.
It is Stable Sorting Algorithm.
It can be useful when number of elements are very less.
It can be also useful when elements are in almost sorted order.

5 4 10 I 6 2

{rt¥Unsorted_
temp

/
1.) 5 10 I 6 2

j
jj
j - -

☐ s to I 6 2

jtl
are [j+☐= temp ;
are [0 ] = 4;

2) 4
g. / to
I 6 2

are [j ] - area]
temp
←

µ
3) 4 5 10 I 6 2
;

"" " "" "" "

" ""
£

☒µ ?
4 ' 62

js
=o
,

if it runs until
j -0 ,
means

ith
correct
position of current

4
5110 0ᵗʰ
6 2 element is at index .

↳ i
j ☐ → represents that current
jᵗʰ index

j
I
4
11
5 ?
,
6 2
before decreament
be same as
and

previous .
its value will

/
I 4 5 10 6 2

jtl
4) EEP
/
I 4 5 10

ji
4
1102
I 5
U
j
I 4 5 ☒ 10 2

jtl
5.)
/
4 56

1g
1 ? ?

1 4 56

j ☐#
4
6110
I 5 ☐
U
j

4 ☐
/
I 5 6 10
↳
j

6110
I ☐ 4 5
or

I 2 4 5 6 10

Insertion Sort is useful in Best Case Only : TC = O(N) {Best Case}

Total Comparisons : N-1 in Best Case
If Array is almost Sorted Insertion Sort will take O(N)
No of swaps = No of Inversions.
Worst Case of Insertion Sort can be when elements are in Descending Order.
Total Comparisons,swaps in Worst Case = 1+2+3+4+5+……..n-1 = O(n²)
Average Case (Half Ascending + Half Descending) = n/2 BC + n/2 WC = O(n²)

https://www.linkedin.com/in/kapilyadav22/
Practice Questions:

Q 1.) No of swaps performed by insertion sort in the given array:

1 2 3 5 0 4 10 8 7

Ans.) 8 swaps.

Q 2.) No of comparisons required to perform insertion sort

25 75 95 125 80 5 10
Ans.) 17 comparisons.
 Selection Sort:
Selection Sort is another Comparison based Sorting Algorithm.
It also uses the concept of Sorted and Unsorted Subarray.
At each pass, we sort one element by finding minimum element from unsorted array
It is Not Stable as the sequence of repeating element can be change.
It is Inplace as it doesn’t require any extra space.

7 4 10 8 3 I
0 I 2 3 4 5

}
7-
4 to 8 3
Pass 1 : -

. 11

/ ↑
° ' 2 3 4 5

sorted unsorted min

3
41min
Pass 2: -
I 4- to 8 3- 7
0 I 2 3 4 5

¥
4 10
Min
Pass 3 :
-
I 3 10 8 41 #
0 I 2 3 4 5

→ -

7 8
Pass 4 : -
I 3 4 81 10
* -

min
0 I 2 3 4 5

1-
& 10

Pass 5 : -
I 3 4 7 10 $

Is
Selection Sort Requires N-1 Passes.
Total Comparisons = ( (N-1) * N ) /2 = O(n²) { Every Case},
Total Swaps = N-1 { Every Case}
Time Complexity = O(n²) {because comparisons are always n².}
Selection Sort speciality is No of swaps are N-1 in every case.
 Quick Sort:
It is a Divide and Conquer Algorithm It is also a type of comparison based Sorting Algorithm.
The Partition Algorithm is the heart of Quick Sort Algorithm.
We choose an element as pivot and perform partition algorithm to the pivot algorithm.
The idea to put pivot element at it’s correct position, which means the left side elements of pivot
elements will be less than or equal to pivot element and the right side elements will be greater than
pivot element.
NOTE: Left and Right side elements need not to be in Sorted Order, to Sort them, we again call Quick
Sort for both the partitions.
It is considered to be inplace algorithm as it only used recursive space.
It is not Stable.

Pivot Selection can be done in Many ways:

Always Select first element as pivot
Always Select last element as pivot
Pick Random element as pivot
Pick Median as pivot.

Let’s choose first element as pivot

pivot
3 9 6 I 5 4 2

pivot pivot
2 I 9 5 4 6

I { } { 6 5 43 { }

pi¥
{ 4,53 E }
/

{ } {5 }
1.) When first element is selected as pivot:

Take i = p = index of first element.

Take second element index as j.
Check if a[j]<= a[p], if yes, increament i, and swap a[j] & a[i].
At the end of the pass swap a[i] and a[p].

2. When last element is selected as pivot:

Take i = index of first element.

Take p = index of last element.
Take j = index of first element, run j loop from j=0 to n-1.
if a[i]<=a[p], swap a[i] and a[j], and then increment i by 1.
At the end, swap a[i] and a[p].

3 9 6 I 5 4 2

{ I } { 6 3 5 49 }

{6 3 54 } { }

{ 3 } { 5,6 }

{ 5} { }
Time Complexity of Quick Sort:

T(n) = T(K-p) + T(q-m) + n + O(1),

where K is the pivot element index after partition, p is the starting index, q is the
last index, O(n) is the time required for partition algorithm, O(1) for choosing pivot
element.

Best Case = T(n/2) + T(n/2) + n, when after partition, there will be equal elements
on both sides.
Worst Case + T(n-1) + n, when after partition, the elements are skewed on one
side only.
Merge Sort:
It is another comparison based sorting algorithm.
It is based on the Divide and conquer technique.
It is Outplace algorithm as we need an additional array to store the temp
output of the sorted array.
It is Stable Sort.

29 1 5 43

29 1 54 3

29 1 54 3

"

2 g I 5 4 3

29 ¥ 45 3

1 29 345

1234 59
 → ◦
G)
→ 1- Cnn)
→ 1- ( n / 2)
◦ (n )

Time Complexity of Merge algorithm is : O(N) for outplace,

O(n/2*n/2) = O(n2) for inplace.

Recurence Relation :
1.) for Outplace:

)={ 047+21-6-2)+n
047

}
Tin it -
=
'

if n > I

TC =
Ocnlogn )
 27 for in place :

÷?
"
"" "

(1) +2T¥
)+n%
◦ n > 1

TC = 042) .

Space complexity : -

input + extra

n
↓
N
logn n

=
Ofn ) .

→
Merge sort is advisable for larger size arrays for .

is
smaller
size ,
Insertion sort
preferable .

Merge Algorithm worst case in

outpace : _

10 20 30 40 11 21 31 41
←
× → for sorted unequal ←e#sy→

=3
→
for sorted
equal -2%
→

elements

comparisons Moves

④ 4) =
4+4-1 ( 4,4 ) =
4+4

(3) %) n I
= -

C. g) =
✗ +
y -1
(E) ± ) -
n

(x ,y)
✗ +
g
TC -
Max ( comparisons ,
moves ]
=
n ( for equal sorted elements )

( for elements)
=

Xty unequal sorted

?⃝
② Merge Algorithm Best case
outpace : -

Yo 50 60 70 do 10 20 30

← →
✗ → ←
y

Comparison Monies

y)
⇐y) mink , ⇐ g) y
= = ✗+

⇐ E) : ⇐ E) n
= =
 Heap Sort:

Heap Sort is a comparison-based Sorting Technique.

It is based on Binary Heap.
Heap sort is an in-place algorithm.
It is not stable, because the heap operations can change the relative ordering of the elements.
We use heapify method in Heap Sort.
It is similar to selection sort, but we can maximum element from heap in constant time.

1. If Complete Binary tree contains K-levels, Total Nodes =

E- I
2. Total leaf nodes in CBT =
In / 27 @ -
total nodes)
3. Total Internal nodes in CBT = =

↳ 12J

Representation of Binary Tree:

1.Array
2. Linked List

NOTE: If node is stored in 1st place of the array then

Away starting from i.

1.) Parent of i = Li /2)

2.) Left child of i = 2 * i
3.) Right child of i = (2*1) + 1
} {different when index changes }

NOTE: Array is better for complete and almost complete Binary Tree.
Linked List is better idea for any gap in Nodes.

⑤\⑧_②_⑧
for example : -

Using Array
: -

D. i, A C B
O 1 2 34 5 6 78 9 10 11 121314 IS

② Using Linked List : -

\ D 20 ]
100 A 300

200 \ C 400
I
300 \ B \

400
?⃝
 MaxHeap: Root is maximum or equal to children at every level.
for ith max, Total comparison = i(i-1)/2 = O(i²).

In Bubble Sort, every pass is costly O(n), but in heap sort, every comparison cost is O(1).

Insertion in MaxHeap (Bottom to top Approach):

6° 35
70
70 So 40 453/539 16 to 9 &
I 2 3 4 5 6 7 8 9 10

s/o 60
40
Liz]
5
parent of 10 = -
.

[
60 50
7° $ 40 456/039 16 10 9 35
US
39 16 5
I 2 3 4 6 7 8 9 10

it 5=[5-2]--2
parent of

to 9
-

60/35 70 60 40 45 SO 39 16 10 9 35

I 2 3 4 5 6 7 8 9 10

parent of 2 is
(2-2)--1
=
.
 and element into
NOTE
Inserting Deleting maxheqp
:-
an ,

will
which
already contains n elements take
order
of login .

Deletion in Max Heap: (Top to Bottom Approach):

50-10-45
%✓ys3533162S3412t@
I 2 3 4 5 6 7 8 9

33
-4s
35

12

3-3
"

zg 3. y
45 10 3533 16 2534
12$50
, g I 2 3 4 5 6 7 8 9

10

1-2
§ 45
I
3335
2 3
10

4
16 25
5 6
34
7
12

8
So

45 3335 12 16 25 34 10 So

I 2 3 4 5 6 7 8 9

NOTE: To delete an element , we can simply put the deleted element at last place, and reduce the loop to one place (from 1-8).
As soon as we deleted the element, we will get the elements in ascending order ( because we are placing max element
at last place.)
 Heap Sort: I 2 3 4 5
A IS 20 7- 9
/ 30

IS Insertion
20
2° 20
→ Insertion Insertion

→
→

IS 15 7 15 7-

30 20 15

%etim
→
Dg
20 15 7- 9
, 7

9 15 9

9
7 Deletion Deletion +
Deletion
→

I
→

7
g
For Insertion in a max heap or min heap, O(logn) times requires for an single element.
for n elements = O(nlogn).
for Deletion in a max or min heap, O(logn) times require for an single element.
for deletion of n elements = O(nlogn)
For heap sort, creation and deletion requires for n elements = O(2nlogn).

NOTE: This method takes O(nlogn) to create a max heap, but there is another method, which is heapify
method and it takes O(n) to create a Build Heap.

Heapify Method: We apply Heapify Method for non-Leaf Nodes.

⑥
30
is

$ 17 20/30/15/20

3012%
to

Heapify Method takes O(logn) time.

 Build Max
↓ heap
→ ocn )

} Deletion

TC =
0cm )
↓
+
Ocnlogn )
Build Max heap