6.

Sorting: Iterative Algorithms

6.1 Selection Sort

6.2 Bubble Sort (covered in Tutorial 2)
6.3 Insertion Sort
6.1 Selection Sort

89 45 68 90 34 17
17 | 45 68 90 34 89
17 34 | 68 90 45 89
17 34 45 | 90 68 89
17 34 45 68 | 90 89
17 34 45 68 89 | 90

- Search the array for the smallest element and swap that element with the
first one.
- Search for the second smallest and swap that element with the second one.
- Repeat the described steps until the whole array is sorted.
6.1 Selection Sort
ALGORITHM selectionSort(A[0..n-1])
for i  0 to n-2 do

find the smallest element A[mini] among

A[i],…,A[n-1]

swap A[i] and A[mini]

- Search the array for the smallest element and swap that element with the
first one.
- Search for the second smallest and swap that element with the second one.
- Repeat the described steps until the whole array is sorted.
6.1 Selection Sort
ALGORITHM selectionSort(A[0..n-1])
for i  0 to n-2 do main operation
mini  i (find the smallest
for j  i+1 to n-1 do element A[mini]
if A[j] < A[mini] mini  j among
A[i],…,A[n-1])
tmp  A[i]
(swap A[i] and
A[i]  A[mini] A[mini])
A[mini]  tmp

- Search the array for the smallest element and swap that element with the
first one.
- Search for the second smallest and swap that element with the second one.
- Repeat the described steps until the whole array is sorted.
i j A[0] A[1] A[2] A[3] A[4] A[5] mini
0 89 45 68 90 34 17 0
0 1 89 45 68 90 34 17 1
0 2 89 45 68 90 34 17 1 find the smallest element i=0:
0 3 89 45 68 90 34 17 1 among A[0], A[1], …, A[5] n-1 comparisons
0 4 89 45 68 90 34 17 4
0 5 89 45 68 90 34 17 5
0 5 17 45 68 90 34 89 5
1 17 45 68 90 34 89 1
1 2 17 45 68 90 34 89 1 find the smallest element
1 3 17 45 68 90 34 89 4 i=1:
among A[1],…, A[5] n-2 comparisons
1 4 17 45 68 90 34 89 4
1 5 17 45 68 90 34 89 4
1 5 17 34 68 90 45 89 4
2 17 34 68 90 45 89 2
2 3 17 34 68 90 45 89 2 find the smallest i=2:
2 4 17 34 68 90 45 89 4 among A[2],…, A[5] n-3 comparisons
2 5 17 34 68 90 45 89 4
2 5 17 34 45 90 68 89 4
3 17 34 45 90 68 89 3
3 4 17 34 45 90 68 89 4  
3 5 17 34 45 90 68 89 4
3 5 17 34 45 68 90 89 4
4 17 34 45 68 90 89 4 find the smallest i=n-2:
4 5 17 34 45 68 90 89 5 among A[4], A[5] 1 comparison
4 5 17 34 45 68 89 90 5
Total: (n−1) + (n−2) + … + 2 + 1 = ½n(n-1) comparisons. The running time is O(n2).
 6.1 Selection Sort: Proof of Correctness
ALGORITHM selectionSort(A[0..n-1])
for i  0 to n-2 do
mini  i
for j  i+1 to n-1 do
if A[j] < A[mini] mini  j
swap A[i] and A[mini]

▪ The array at any moment is divided into two parts: sorted and unsorted, and
any element of the sorted sub-array is less than or equal to any element of
the unsorted one.

▪ In each pass of the algorithm, the minimum element of the unsorted sub-
array is found and swapped with the first element of the unsorted sub-array.
The sorted sub-array is thus extended by one element in every pass.
 6.1 Selection Sort: Proof of Correctness
ALGORITHM selectionSort(A[0..n-1])
for i  0 to n-2 do
mini  i
for j  i+1 to n-1 do
if A[j] < A[mini] mini  j
swap A[i] and A[mini]
Loop invariant

At the beginning of iteration i, the sub-array A[0..i-1] is sorted and

any element in A[0..i-1] is less than or equal to any element in
A[i..n-1].

1) The invariant is true initially: the sub-array A[0..-1] is empty.

2) If the invariant is true for i=k, it is true for i=k+1.
3) When the loop terminates, the invariant is true for the entire array.
4) The loop does terminate.
6.2 Bubble Sort
ALGORITHM bubbleSort(A[0..n-1])
for i  0 to n-2 do main operation
for j  0 to n-2-i do
if A[j]>A[j+1]
tmp  A[j]
A[j]  A[j+1]
A[j+1]  tmp

Tracing and analysis of Bubble sort was done in Tutorial 2.

O(n2)
6.3 Insertion Sort

i A[0] A[1] A[2] A[3] A[4] A[5]

1 23 78 45 8 32 56
2 23 78 45 8 32 56
3 23 45 78 8 32 56
4 8 23 45 78 32 56
5 8 23 32 45 78 56
6 8 23 32 45 56 78

The array at any moment is divided into two parts: sorted and unsorted.
In each pass of the algorithm, the first element of the unsorted sub-list
is inserted into the sorted sub-list at the appropriate place.
 6.3 Insertion Sort

ALGORITHM insertionSort(A[0..n-1])
//The algorithm sorts a given array A
//Input: an array A[0..n-1] of n numbers
//Output: array A[0..n-1] in ascending order
for i  0 to n-1 do main operation
tmp  A[i]
j i-1
while (j≥0 and tmp < A[j]) do
A[j+1]  A[j]
j  j-1
A[j+1]  tmp

Analysis: CW1
 Testing the Sorting Algorithms

n Selection Sort Insertion Sort Bubble Sort Merge-Sort Quick Sort

10,000 5.17 sec. 3.18 sec. 10.88 s. 0.05 sec. 0.05 sec.

20,000 22.03 sec 13.51 sec. 45.09 s. 0.11 sec 0.05 sec.

40,000 1 min. 31.01 sec. 53.66 sec 2 min. 59.40 s. 0.22 sec. 0.11 sec.

80,000 6 min. 06.30 sec. 3 min. 46.62 sec. 12 min. 27.15 s. 0.49 sec. 0.28 sec.

iterative recursive

from A. Drozdek “Data Structures and Algorithms in C++”

(Pentium 133 MHz PC)