Advanced Programming

Sorting Algorithms

Sorting - Iterative Algorithms

Problem Definition
Input: A sequence of n elements <a1, a2, , an> Output: A permutation <a1, a2, , an> of such elements, so that a1 a2 an

Advanced Programming

Sorting Algorithms

Types of Ordering
Internal

Ordering

All the elements to be ordered are in main memory Direct access to all elements
External

Ordering

Elements cannot be loaded all in memory at the same time It is necessary to act on elements stored on a file Usually, sequential access
3

Practical observations
Elements
The

to be ordered are usually structures (struct) made of many variables (fields)

key of such structure is usually one field (or a value calculated from one or more fields) Remaining fields are additional data but useless for ordering Ordering is made for increasing values of the key
4

Advanced Programming

Sorting Algorithms

Example
struct student { int id; char surname[30] ; char name[30] ; int grade; } ;
struct student class[100] ;
5

Example
struct student { int id; char surname[30] ; Ordering by name char name[30] ; and surname (key = concatenation int grade; name and surname) } ; Ordering by grade(
repeated values) Ordering by id

struct student class[100] ;

Advanced Programming

Sorting Algorithms

Stability

A sorting algorithm is called stable whenever, even if there are elements with the same value of the key, in the resulting sequence such elements appear in the same order in which they appeared in the initial sequence.

Simple Assumption

During the study of sorting algorithms there are usually arrays of n integer values: int A[n] ;

Advanced Programming

Sorting Algorithms

Algorithms
There are many sorting algorithms with different complexity: O(n2): simple, iterative

Insertion sort, Selection sort, Bubble sort,

O(n):

only applicable in particular cases

Counting sort, Radix sort, Bin (or Bucket) sort,

O(n

log n): more complex, recursive

Merge sort, Quicksort, Heapsort

Insertion sort
Already sorted v[j] 2 3 6 12 16 21 8 Move forward all the elements so that v[ I ] > v[ j ] 2 3 6 8 12 16 21 Not considered yet

12 16 21

5
10

Advanced Programming

Sorting Algorithms

Pseudo-code

Insert A[ j ] in ordered sequence A [1 .. j-1]

11

Implementation in C
void InsertionSort(int A[], int n) { int i, j, key ; for(j=1; j<n; j++) { key = A[j] ; i = j - 1 ; while ( i >= 0 && A[i]>key ) { A[i+1] = A[i] ; i-- ; } A[i+1] = key ; } }
12

Advanced Programming

Sorting Algorithms

From pseudo-code to C
Note well: In C, array indexes are from 0 to n-1, while pseudo-code use ranges from 1 to n. Indentation of code is useful but remember braces to identify blocks { }

13

Complexity
Number of comparisons: Cmin = n-1 Cavg = (n2+n-2) Cmax = (n2+n)-1 Number of data-copies Mmin = 2(n-1) Mavg = (n2+9n-10) Mmax = (n2+3n-4) C = O(n2), M = O(n2) T(n) = O(n2) T(n) non Tworst case(n) = (n2) (n2)
14

Best case: array already ordered Worst case: array ordered inversely

Advanced Programming

Sorting Algorithms

Other quadratic algorithms

Insertion Sort Selection Sort Bubble Sort

15

Execution Time (ms)

Random

Inversely Ordered

n = 256

512

256

512

256

512

16

Advanced Programming

Sorting Algorithms

Impact of data
n = 256

2 byte 2 byte 2 byte 16 byte 16 byte 16 byte

17

Counting sort
It cannot be applied in general, as it is based on this hypothesis: The n elements to be ordered are integer numbers between 1 and k, with k integer. With such hypothesis, if k = O(n), then the algorithms complexity is just O(n).

18

Advanced Programming

Sorting Algorithms

Basic Idea

Find out, for each element x, how many elements of the array are less than x. Such information allows to put x directly in the final position in the array

19

Data Structure
Three

arrays are needed:

Initial array: A[1..n] Final array: B[1..n] Temporary Array: C[1..k]

C keeps track of number of elements of A having a certain value: C[i] is the number of elements of A equals to i. Sum of the first i elements of C defines the number of elements of A whose values is <= i
20

10

Advanced Programming

Sorting Algorithms

Pseudo-code

21

Analysis
For each j, C[ A[j] ] represents the number of elements less than or equals to A[j], and then it is the final position of A[j] in B: B[ C[ A[j] ] ] = A[j] The correction C[ A[j] ] C[ A[j] ] 1 is needed to handle duplicate elements.

22

11

Advanced Programming

Sorting Algorithms

Example (n=8, k=6)

23

Example (2)
A B B B B B B B B 1 1 1 1 1 1 1 1 1 3 3 3 6 4 1 3 4 1 4 4 4 4 4 4 4 4 4 C C C C C C C C 2 2 4 6 7 8 1 2 4 6 7 8 1 2 4 5 7 8 1 2 3 5 7 8 0 2 3 5 7 8 0 2 3 4 7 8 0 2 3 4 7 7 0 2 2 4 7 7

j=8
j=7 j=6

j=5 j=4
j=3 j=2 j=1
24

3 4 4 4 3 4 4 4 6

1 1 3 3 4 4 4 6

12

Advanced Programming

Sorting Algorithms

Complexity
Initialization of C: O(k) 3-4: Calculate C: O(n) 6-7: Sum in C: O(k) 9-11: Copy in B: O(n) Total complexity is O(n+k). Algorithm is useful only when k=O(n), because the resulting complexity is O(n).
1-2:

25

Note
The condition of applicability of the algorithm can be extended in this way: The key field of n elements to be ordered has a limited number of possible values k.

26

13

Advanced Programming

Sorting Algorithms

Bubble Sort
In

each cycle compare every couple of consecutive elements and if they are not ordered, then swap (exchange) them. Repeat this process N times and all the elements will be ordered Complexity is O(n2) Optimization: if during last cycle there are no swaps, then the elements are already sorted

27

Bubble sort in C
void BubbleSort(int A[], int n) { int i, j, t; for(i=1; i<n-1; i++) { for(j=1; j<n-1; j++) { if ( A[j]>A[j+1] ) { t = A[j] ; A[j] = A[j+1]; A[j+1] = t; } } } }
28

14

Advanced Programming

Sorting Algorithms

Bubble sort (optimized) in C

15