1
Algorithm Analysis
2
Question Suppose you have 2 programs that
will sort a list of student records and allow you to search for student information.
Question: How do you judge which is a better program?Answer: By running programs or Algorithm Analysis
3
Algorithm Analysis Is a methodology for estimating
the resource (time and space) consumption of an algorithm.
Allows us to compare the relative costs of two or more algorithms for solving the same problem.
4
How to Measure “Betterness” Critical resources in computer
Time Memory space
For most algorithms, running time depends on size n of data.
Notation: T(n) Time T is a function of data size n.
5
Two approaches Two approaches to obtaining
running time: Measuring under standard
benchmark conditions (running programs)
Estimating the algorithms performance (analyzing algorithms)
6
Estimation Assumptions Estimation of running time is based
on: Size of the input Number of basic operations
The time to complete a basic operation does not depend on the value of its operands.
7
Example: largest()
// Return position of largest value// in array
int largest(int array[], int n) { int posBig = 0; // 1 for (int i = 1; i < n; i++)
if (array[posBig] < array[i]){ // 2 posBig = i; // 3 } return posBig; }
8
Example: largest() Step 1: T(n) = c1
Step 2: It takes a fixed amount of time to do one comparison T(n) = c2n
Step 3: T(n) = c3n, at most Total time: T(n) = c1 + c2n+ c3n
≈ c1 + (c2 + c3)n ≈ cn
9
Example: Assignment
int a = array[0];
The running time is T(n) = c This is called constant running time.
10
Example: total()
sum = 0;for (i = 1; i <= n; i++) for (j = 1; j < n; j++) sum++;} What is the running time for this code? Analysis:
The basic operation is sum++ The cost of a sum operation can be bundled into
some constant time c. Inner loop takes c1n; outer loop takes c2n. The total running time is T(n) = c1n * c2n
= (c1 + c2)n = c3n2
11
Growth Rate of Algorithm—Big-O Natation Growth Rate for an algorithm
is the rate at which the cost of the algorithm grows as the data size n grows.
Constant: T(n) = c O(1) Linear Growth: T(n) = n O(n) Quadratic Growth: T(n) = n2 O(n2) Exponential Growth: T(n) = 2n O(2n)
Assumptions Growth rates are estimates. They have meaning when n is large.
Growth Rate Graph
12
13
Growth Rate Graph
n
T(n)
c
Log n
14
Characteristics of Growth Rates
Growth Rates Constant— T(n) = c
Independent of n Linear— T(n) = cn
Constant slope Logarithmic— T(n) = c log n
Slope flattens out quickly Quadratic— T(n) = cn2
Increasing slope Cubic— T(n) = cn3
Exponential— T(n) = c2n
Growth Rate—Array Operations
15
bool search(int list[], int n, int key) found = false Loop (for i = 0 to i < n) If (list[i] == key) Then found = true End If End loop return found End // search
Growth Rate—Array Operations
16
• Linear Search• Insert at Back• Insert at Front• Remove from Back• Remove from Font
Linear Search
17
bool search(int list[], int n, int key) found = false // c1 Loop (for i = 0 to i < n) If (list[i] == key) Then // nc2
found = true (at most) End If End loop return found // c3
End // searchT(n) = c1 + nc2 + c3 = c + nc2 ≈ nc2 ∴ O(n)
Insert at Back
18
void insertAtBack(int list[], int &n, item) If (!isFull) Then list[n] = item n++ End IfEnd // insertAtBack
Insert at Back
19
void insertAtBack(int list[], int &n, item) If (!isFull) Then list[n] = item // c1 n++ // c2 End IfEnd // insertAtBack
T(n) = c1 + c2 = c ∴ O(1)
Insert At Front
20
void insertAtFront(int list[], int &n, int item) If (!isFull) Then // shift to right Loop (for i = n downto 1) list[i] = list[i – 1] End loop // assign list[0] = item count++ End IfEnd // insertAtFront
Insert At Front—Growth Rate?
21
void insertAtFront(int list[], int &n, int item) If (!isFull) Then Loop (for i = n downto 1) list[i] = list[i – 1] // c1n (at most) End loop
list[0] = item // c2
count++ // c3
End IfEnd // insertAtFront
T(n) = c1n + c2 + c3 = nc1 + c ≈ nc1 ∴ O(n)
Remove From Back
22
int removeFromBack(list[], int &n) result = NIL If (!isEmpty) Then result = list[n – 1] n = n - 1 End If return resultEnd // removeFromBack
Remove From Front
23
int removeFromFront(list[], int &n) result = NIL If (!isEmpty) Then result = list[0] // shift left Loop (for i = 0 to n – 2) list[i] = list[i + 1] End Loop n = n - 1 End If return resultEnd // removeFromFront
Growth Rate—Linked-List Operations
24
• Linear Search• Insert at Back• Insert at Front• Remove from Back• Remove from Font
Linear Search
25
bool linearSearch(int key){ bool found = false; c1
if (head != NULL){ node* temp = head; c2
while (temp != NULL){ if (temp=>getData() != key) found = true; else temp = temp->getNext(); c3n } } return found;} T(n) = c1 + c2 + c3n ≈ cn ∴ Growth Rate: O(n)
insertAtBack()
26
void insertAtBack(int item){ node *newP = new node(item, NULL); node *temp = head; c1
if (head == NULL) head = newP; else while (temp->getNext() != NULL){ temp = temp->getNext(); c2n } temp->setNext(newP); c3
count++; c4
} T(n) = c1 + c2n + c3 + c4 ≈ c + c2n ∴ G.R. = O(n)
insertAtFront()
27
void insertAtFront(int item){ node *newP = new node(item, NULL); c1
newP->setNext(head); c2
head = newP; Next(newP); c3
count++; c4
}
T(n) = c1 + c2 + c3 + c3 = c ∴ G.R. = O(1)
removeFromBack()
28
void removeFromBack(){ node *temp = head; c1
node *tail = NULL; c2
while (temp != NULL){
trail = temp; nc3
temp = temp->getNext(); nc4 } count++; c5
} T(n) = c1 + c2 + nc3 + nc4 + c5 = c6 + n(c3 + c4) ≈ cn ∴ G.R. = O(n)
removeFromFront()
29
void removeFromBack(){ node *newP = new node(item, NULL); c1
newP->setNext(head); c2
head = newP; Next(newP); c3
count++; c4
} T(n) = c1 + c2 + c3 + c3 = c ∴ G.R. = O(1)
30
Best, Worst, Average Cases
For an algorithm with a given growth rate, we consider Best case Worst case Average case
For example:
Sequential search for K in an array of n integers
Best case: The first item of the array equals K Worst case: The last position of the array equals K Average case: Match at n/2
31
Which Analysis to Use The Worst Case.
Useful in many real-time applications. Advantage:
PredictabilityYou know for certain that the algorithm must perform at least that well.
Disadvantage: Might not be a representative measure of
the behavior of the algorithm on inputs of size n.
32
Which Analysis to Use? The average case.
Often we prefer to know the average-case running time.
The average case reveals the typical behavior of the algorithm on inputs of size n.
Average case estimation is not always possible.
For the sequential search example, it assumes that the key value K is equally likely to appear in any position of the array. This assumption is not always correct.
33
Which Analysis to Use The Best Case
Normally, we are not interested in the best case, because:
It is too optimistic Not a fair characterization of the
algorithms’ running time Useful in some rare cases, where the best
case has high probability of occurring.
34
The moral of the story If we know enough about the
distribution of our input we prefer the average-case analysis.
If we do not know the distribution, then we must resort to worst-case analysis.
For real-time applications, the worst-case analysis is the preferred method.
Your Turn (Stack w/array)
35
Given Stack Class:
…private: elemType data[MAX]; int top;}
Stack (w/array)
36
What is the growth rate (big O) of the push operation?
void push(elemType item){ if (!isFull()){ top++; data[top] = item; }}
Stack (w/array)
37
What is the growth rate (big O) of the pop operation?
elemType pop (){ elemType result = NIL; if (!isEmpty()){ result = data[top]; top--; }}
Stack (w/array)
38
What is the growth rate (big O) of the clear operation?
void clear(){ top = -1;}
Stack (w/linked list)
39
Given: Stack class with linked list
…private: Node *top;}
Stack (w/linked list)
40
What is the growth rate (big O) of the push operation?
void push(elemType item){ Node *temp = new Node(item, NULL); top->setNext(temp); top = temp;}
Stack (w/linked list)
41
What is the growth rate (big O) of the pop operation?
elemType pop () { elemType result = NIL; Node *temp = top; if (!isEmpty()){ result = top-<getData(); top = top->getNext(); delete temp; } return result;}
Stack (w/linked list)
42
What is the growth rate (big O) of the clear operation?
Void clear () { Node *temp = top; while (top != NULL) { top = top->getNext(); delete temp; temp = top; }}
Growth Rate of Binary Search (w/array)
43
Comparisons Number left
0 n1 n(1/2)2 n(1/2)2
3 n(1/2)3
4 n(1/2)4
… …k-1 n(1/2)k-1
k n(1/2)k
Solution n(1k) nn(1/2)k = ------- = ---- = 1 (2k) (2k)2k = n log(2k) = log(n)k log(2) = log(n)k = log(n)/log(2)k = c1 log(n)T(n) = c1 c2 log(n)O(log n)
Analysis of Bubble Sort
void bubbleSort(int a[], int count) pass count – 1 last pass – 1 Loop (for i from 1 to pass) Loop (for j from 0 to last) If (a[j] > a[j + 1]) Then swap a[j] and a[j = 1] End If End Loop last last – 1 End Loop
• Suppose: count = 10maximum pass = 9
pass comparisons1 92 83 74 65 56 47 38 29 1
Analysis of Bubble Sort
• Total number of comparisons 1 + 2 + 3 + 4 + 5 + 6 + 7 + 8 + 9 = ? 9 + 8 + 7 + 6 + 5 + 4 + 3 + 2 + 1 = ?--------------------------------------------------10+10 +10+……………………+10 +10 = 9x10
• (n-1) + (n-2) + (n-3)… + 3 + 2 + 1 1 + 2 + 3 …+ (n-3) + (n-2) + (n-1) n + n + n … + n + n + n
• = n(n-1)• Therefore, ? = n(n – 1)/2
= cn2 – cn ≈ cn2
• Growth Rate: O(n2)