Algorithms Analysis

Introduction
 An algorithm is a clearly specified set of simple instructions to be followed to
solve a problem.

 Once an algorithm is given for a problem and decided (somehow) to be correct,

an important step is to determine how much in the way of resources, such as
time or space, the algorithm will require.

 An algorithm that solves a problem but requires a year is hardly of any use.
Likewise, an algorithm that requires hundreds of gigabytes of main memory is not
(currently) useful on most machines.

 To evaluate an algorithm’s efficiency, real-time units such as microseconds and

nanoseconds should not be used. Rather, logical units that express a relationship
between the size n of a file or an array and the amount of time t required to
process the data should be used.
2
Machine Independence
 The evaluation of efficiency should be as machine independent as possible.

 It is not useful to measure how fast the algorithm runs as this depends on
which particular computer, OS, programming language, compiler, and kind
of inputs are used in testing

 Instead,

 we count the number of basic operations the algorithm performs.

 we calculate how this number depends on the size of the input.

 A basic operation is an operation which takes a constant amount of time to

execute.

 Hence, the efficiency of an algorithm is the number of basic operations it

performs. This number is a function of the input size n.
3
Example of Basic Operations
 Arithmetic operations: *, /, %, +, -

 Assignment statements of simple data types

 Reading of primitive types

 Writing of a primitive types

 Simple conditional tests: if (x < 12) ...

 Method call (Note: the execution time of the method itself may depend on the value of
parameter and it may not be constant)

 A method's return statement

 Memory Access

 We consider an operation such as ++ , += , and *= as consisting of two basic

operations.

4
 Note: To simplify complexity analysis we will not consider memory access (fetch or
store) operations.

Problem Size
 For every algorithm we want to analyze, we need to define the size of the
problem
 For a search algorithm, the size of the problem is the size of the search pool
 For a sorting algorithm, the size of the program is the number of elements
to be sorted
 The efficiency of an algorithm is always stated as a function of the problem
size
 We generally use the variable n to represent the problem size

5
Problem/Input size matters!
Some example algorithms and their expected running times based on
the input size

6
Growth Functions

 We must also decide what we are trying to efficiently optimize

 time complexity – CPU time

 space complexity – memory space

 CPU time is generally the focus

 The rates of growth are expressed as functions, which are generally

in terms of the number of inputs n

 A growth function shows the relationship between the size of the

problem (n) and the value we hope to optimize (time). This function
represents the time complexity or space complexity of the algorithm.

7
Algorithm Complexity

 Worst Case Complexity:

 The function defined by the maximum number of steps taken on any
instance of size n

 Best Case Complexity:

 The function defined by the minimum number of steps taken on any
instance of size n

 Average Case Complexity:

 The function defined by the average number of steps taken on any
instance of size n

8
 Algorithm Complexity (cont…)
 We are usually interested in the worst case complexity: what are the most
operations that might be performed for a given problem size.

 Best case depends on the input

 Average case is difficult to compute

 So we usually focus on worst case

analysis
 Easier to compute

 Usually close to the actual running

time

 Crucial to real-time systems

9
Algorithm Complexity (cont…)
 Example: Linear Search Complexity

 Best Case: Item found at the beginning: One comparison

 Worst Case: Item found at the end: n comparisons

 Average Case: Item may be found at index 0, or 1, or 2, . . . or n - 1

Average number of comparisons is: (1 + 2 + . . . + n) / n = (n + 1) / 2
Worst and Average complexities of common sorting algorithms:
Method Worst Case Average Case

Selection sort n2 n2

Inserstion sort n2 n2

Merge sort n log2 n n log2 n

Quick sort n2 n log2 n

10
Running Time Analysis

11
Asymptotic Complexity

 It is not typically necessary to know the exact growth function for an

algorithm

 Finding the exact complexity, f(n) = number of basic operations, of

an algorithm is difficult

 We are mainly interested in the asymptotic complexity of an

algorithm – the general nature of the algorithm as n increases

 Asymptotic complexity is based on the dominant term of the growth

function – the term that increases most quickly as n increases

12
Asymptotic Complexity (cont…)
 We approximate f(n) by a function g(n) in a way that does not substantially
change the magnitude of f(n). The function g(n) is sufficiently close to f(n)
for large values of the input size n.

 This "approximate" measure of efficiency is called asymptotic complexity.

 Thus the asymptotic complexity measure does not give the exact number
of operations of an algorithm, but it shows how that number grows with the
size of the input.

 This gives us a measure that will work for different operating systems,
compilers and CPUs.

 Asymptotic bounds are used to estimate the efficiency of algorithms by

assessing the amount of time and memory needed to accomplish the task for
which the algorithms were designed.

13
 Notations
 Following are the commonly used asymptotic notations to calculate the running time
complexity of an algorithm.
 Ο Notation
 Ω Notation
 θ Notation

 O(expression) gives an upper bound on the growth rate of a function. It specifically

describes the worst-case scenario or the longest amount of time an algorithm can
possibly take to complete.

 Omega(expression) gives a lower bound on the growth rate of a function. It measures

the best case time complexity or the best amount of time an algorithm can possibly
take to complete.

14
 Theta(expression) consist of all the functions that lie in both O(expression) and
Omega(expression). When the upper and lower bounds are the same within a
constant factor, we indicate this by using θ(big-Theta) notation.

Meanings of the various growth functions

15
Big-Oh Notation
 The most commonly used notation for specifying asymptotic complexity is the big-O
notation.

 The coefficients and the lower order terms become increasingly less relevant as n
increases
 So we say that the algorithm is order n2, which is written O(n2)

 This is called Big-Oh notation

 There are various Big-Oh categories

 Two algorithms in the same category are generally considered to have the same
efficiency, but that doesn't mean they have equal growth functions or behave exactly
the same for all values of n

 The Big-O notation, O(g(n)), is used to give an upper bound (worst-case) on a

positive runtime function f(n) where n is the input size.

16
 O Notation: Definition
 Big O:
 T(n) = O(f(n)) if there are positive constants c and N
such that T(n) ≤ c f(n) when n ≥ N
 This says that function T(n) grows at a rate no faster than f(n) ; thus f(n)
is an upper bound on T(n).

 Another way:
f(n) is O(g(n))↔ there exist numbers c, N > 0
such that for each n ≥ N
f(n) ≤ c.g(n)
The meaning:
• f(n) is larger than g(n) only for finite number of n’s;
• a constant c and a value N can be found so that for every value of n ≥ N: f(n) ≤ c.g(n);
• f(n) does not grow more than a constant factor faster than g(n).

17
O Notation: illustration

c. g(n)

f(n )

N n T(n) = O(f(n))

18
Big-Ω Notation: Definition

 Big Omega:
 T(n) = Ω(f(n)) if there are positive constants c
and N such that T(n) ≥ c f(n) when n ≥ N
 This says that function T(n) grows at a rate no slower than f(n) ;
thus f(n) is a lower bound on T(n).

 Another way:
f(n) is Ω(g(n))↔ there exist numbers c, N > 0
such that for each n ≥ N
f(n) ≥ c.g(n)

19
Big-Ω Notation: illustration

T(n) = Ω(f(n))
Big-θ Notation

 The definitions of big-Oh and Ω allow describing the upper bound

of an algorithm and the lower bound of an algorithm

 When the upper and lower bounds are the same within a
constant factor, we indicate this by using Θ (big-Theta) notation.

 An algorithm is said to be Θ(h(n)) if it is in O(h(n)) and it is in

Ω(h(n)).
Big-θ Notation: Definition
 Big Theta:
 T(n) = θ(f(n)) if and only if

T(n) = O(f(n)) and T(n) = Ω(f(n))

 This says that function T(n) grows at the same rate as f(n)

 Another way:

f(n) is θ(g(n))↔ there exist numbers c1, c2, N > 0

such that for each n ≥ N
c1. g(n) ≤ f(n) ≤ c2. g(n)

big-Ω notation big-O notation

22
Big-θ Notation: illustration

c2 g(n)

f(n )

c1 g(n)

N n
Big-Oh Categories

 Some sample growth functions and their Big-Oh

categories:

24
Rules for using big-O
 For large values of input n, the constants and terms with lower

degree of n are ignored.

1. Multiplicative Constants Rule: Ignoring constant factors.

O(c f(n)) = O(f(n)), where c is a constant;

Example:

O(20 n3) = O(n3)

2. Addition Rule: Ignoring smaller terms.

 If O(f(n)) < O(h(n)), then O(f(n) + h(n)) = O(h(n)).

 i.e. If T1(n) = O(f(n)) and T2(n) = O(g(n)), then

25
 T1(n) + T2(n) = max(O(f(n)), O(g(n)))

Example 1:

O (n2 log n + n3) = O(n3)

O (2000 n3 + 2n! + n800 + 10n + 27n log n + 5) = O(n!)

Example 2 (Algorithm A):

Step 1: Run algorithm A1 that takes O(n3) time

Step 2: Run algorithm A2 that takes O(n2) time

TA(n) = TA1(n) + TA2(n) = O(n3) + O(n2) = max (O(n3), O(n2)) = O(n3)

3. Multiplication Rule:
O(f(n) * h(n)) = O(f(n)) * O(h(n))

i.e. If T1(n) = O(f(n)) and T2(n) = O(g(n)),

then

26
 T1(n) * T2(n) = O(f(n)) * O(g(n))

Example:

O((n3 + 2n2 + 3n log n + 7) (8n2 + 5n + 2)) = O(n5)

4. If T(n) is a polynomial of degree k, then

T(n) = O(nk)

Example:

T(n) = n8 + 3n5 + 4n2 + 6 = O(n8)

logk(n) = O(n) for any constant k

27
Comparing Growth Functions

 You might think that faster processors would make efficient

algorithms less important

 A faster CPU helps, but not relative to the dominant term

28
Comparing Growth Functions (cont…)

A hierarchy of growth rates:

c < log n < log2 n < logk n < n < n log n

< n2 < n3 < 2n < 3n < n! < nn

29
Comparing Growth Functions (cont…)

As n increases, the various growth functions diverge dramatically:

30
Comparing Growth Functions (cont…)

31
How to determine complexity of code structures

NOTE : In general, doing something with every item in one

dimension is linear, doing something with every item in two
dimensions is quadratic, and dividing the working area in half is
logarithmic.

32
Analyzing Loop Execution
 Loops: for, while, and do-while:
 First determine the order of the body of the loop, then multiply that by the
number of times the loop will execute
for (int count = 0; count < n; count++)
// some sequence of O(1) steps

 N loop executions times O(1) operations results in a O(n) efficiency

Consider the following loop:
count = 1; while
(count < n)
{ count *= 2;
// some sequence of O(1) steps
}

 The loop is executed log2n times, so the loop is O(log n)

33
Analyzing Loop Execution (cont…)
 Loops: for, while, and do-while:
Again: complexity is determined by the number of iterations in the
loop times multiplied by the complexity of the body of the loop.

 Examples:

for (int i = 0; i < n; i++)

sum = sum - i; O(n)

for (int i = 0; i < n * n; i++) O(n2)

sum = sum + i;

34
 Analyzing Loop Execution (cont…)
i=1;
while (i < n) { sum
= sum + i;
i = i*2 O(log n)
}

 We start by considering how to count operations in for-loops.

 First of all, we should know the number of iterations of the loop; say it is x.
 Then the loop condition is executed x + 1 times.

 Each of the statements in the loop body is executed x times.

 The loop-index update statement is executed x times.

35
 Analyzing Loop Execution (cont…)
 Example:
Time Units to Compute:
int sum (int n) -----------------------•
1 for the assignment
{ int partial_sum = 0;
• Loop Statement: 1 assignment, n+1 tests, and
int i; n increments
for (i = 1; i <= n; i++) • Loop Body: n loops of 3 units for:
(an assignment, an addition, and multiplications)
partial_sum = partial_sum + (i * i);
• 1 for the return statement
return partial_sum; ------------------------
Total: 1 + (1 + n + 1 + n) + 3n + 1
}
= 5n + 4 = O(n)

36
 Analyzing Loop Execution (cont…)
 Loops (with <):

 In the following for-loop:

for (int i = k; i < n; i = i + m){
statement1;
statement2;
}
 The number of iterations is: (n – k ) / m
 The initialization statement, i = k, is executed one time.

 The condition, i < n, is executed (n – k ) / m + 1 times.

 The update statement, i = i + m, is executed (n – k ) / m times.
 Each of statement 1 and statement 2 is executed (n – k ) / m times.

37
Analyzing Loop Execution (cont…)
 Loops (with <=):

 In the following for-loop:

for (int i = k; i <= n; i = i + m){
statement1;
statement2;
}
 The number of iterations is: (n – k) / m + 1
 The initialization statement, i = k, is executed one time.

 The condition, i <= n, is executed (n – k) / m + 1 + 1 times.

 The update statement, i = i + m, is executed (n – k) / m + 1 times.
 Each of statement 1 and statement 2 is executed (n – k) / m + 1 times.
 Loop example:
 Find the exact number of basic operations in the following program fragment:

38
 Analyzing Loop Execution (cont…)
double x, y; x = 2.5 ; y =
3.0; for(int i = 0; i < n;
i++){ a[i] = x * y; x = 2.5
* x; y = y + a[i]; }

 There are 2 assignments outside the loop 2 operations

 The for loop actually comprises

 an assignment (i = 0) 1 operation

 a test (i < n) n + 1 operations

 an increment (i++) 2 n operations

 the loop body that has three assignments, two multiplications, and an addition 6 n
operations
Thus the total number of basic operations is 6 * n + 2 * n + (n + 1) + 3 = 9n + 4

39
Analyzing Loop Execution (cont…)
 Loops With Logarithmic Iterations:

 In the following for-loop: (with <)

for (int i = k; i < n; i = i * m){
statement1;
statement2;
}
 The number of iterations is: (Logm (n - k) )

 In the following for-loop: (with <=)

for (int i = k; i <= n; i = i * m){
statement1;
statement2;
}
 The number of iterations is: (Logm (n - k) + 1)

40
Analyzing Nested Loops:

 When loops are nested, we multiply the complexity of the outer loop
by the complexity of the inner loop
for (int count = 0; count < n; count++)

for (int count2 = 0; count2 < n; count2++)

// some sequence of O(1) steps

 Both the inner and outer loops have complexity of O(n)

 The overall efficiency is O(n2)

 Nested Loops: Complexity of inner loop * complexity of outer loop

41
Analyzing Loop Execution (cont…)
 Examples:
sum = 0
for(int i = 0; i < n; i++)
for(int j = 0; j < n; j++) O(n2)
sum += i * j ;

i = 1;
while(i <= n) { j = 1;
while(j <= n){ statements of constant
complexity; j = j*2;
O(n log n)
}
i = i+1;
}

42
Analyzing Loop Execution (cont…)
Example:

for(int i = 1; i <= n; i++) 3mn = O(mn)

for(int j = 1; j <= m; j++)
sum = sum + i + j;

 Example:
for(int i = 1; i <= n; i++)
for(int j = 1; j <= m; j++)
for(int k = 1; k <= p; k++) sum =
sum + i + j + k; 4pmn = O(pmn)

43
Analyzing Sequence of Statements

 Consecutive statements: Use Addition rule

 These just add, and the maximum is the one that counts

O(s1, s2, s3, … ,sk) = O(s1) + O(s2) + O(s3) + … + O(sk)

= O(max(s1, s2, s3, . . . , sk))
Example:

for (int j = 0; j < n * n; j++)

sum = sum + j;

for (int k = 0; k < n; k++)

sum = sum - l;
System.out.print("sum is now ” + sum);
 Complexity is O(n2) + O(n) + O(1) = O(n2)

44
Analyzing Sequence of Statements (cont…)
 Consecutive statements: Use Addition rule

Example:

for (i = 1; i <= n; i++)

O(n)
sum = sum + i;

for (i = 1; i <= n; i++) {

for (j = 1; j <= n; j++) O(n2)
sum = sum + i + j;
}

 Complexity is O(n2 + + n) = O(n2)

45
Analyzing Sequence of Statements (cont…)
 Consecutive statements: Use Addition rule

Example:

for (i = 1; i <= n; i++)

for (j = 1; j <= n; j++)
O(n2)
sum = sum + i + j;

sum = sum / n; O(1)

for (i = 1; i <= n; i++)
O(n)
sum = sum + i;

for (j = 1; j <= n; j++)

O(n)
sum = sum + j * j;

46
n2 + + 1 + n + n = O(n2 + 2n + 1) = O(n2)

47
48
Analyzing If Execution (cont…)

 if (test) s1 else s2
The running time is never more than the running time of the test plus the
larger of the running times of s1 and s2

 Example:
if (test == 1) O(1)
for (i = 1; i <= n; i++)
sum = sum + i;
O(n)
else
for (i = 1; i <= n; i++)
for (j = 1; j <= n; j++) O(n2)
sum = sum + i + j;

The running time = 1 + max(n, n2) = O(n2)

49
Analyzing If Execution (cont…)
 Sometimes if-else statements must carefully be checked: O(if-else) =
O(Condition) + Max [O(if), O(else)]
int[] integers = new int[n];
........
if(hasPrimes(integers) == true)
integers[0] = 20; O(1)
else
integers[0] = -20; O(1)

public boolean hasPrimes(int[] arr) {

for(int i = 0; i < arr.length; i++)
..........
.......... O(n)
} // End of hasPrimes()

O(if-else) = O(Condition) = O(n)

50
Analyzing If Execution (cont…)
 Example (complexity of a code fragment which include if statement):
for(i = 1; i <= n; i++)
The Running Time:

for(j = 1; j <= n; j++) = O(n3) + O(n2)

= O(n3)
for(k = 1; k <= n; k++)

sum = sum + i + j + k;

if (test == 1)
for (i = 1; i <= n; i++)
for(j = 1; j <= n; j++) sum =
sum + i;
else for (i = 1; i <= n;
i++)
sum = sum + i + j;

51
 Note: Sometimes a loop may cause the if-else rule not to be applicable.
 Consider the following loop:
while (n > 0) { if
(n % 2 == 0) {
System.out.println(n);
n = n / 2;
} else
{
System.out.println(n);
System.out.println(n);
n = n – 1;
}
}
The else-branch has more basic operations; therefore one may conclude
that the loop is O(n). However the if-branch dominates. For example, if n is
60, then the sequence of n is: 60, 30, 15, 14, 7, 6, 3, 2, 1, and 0. Hence
the loop is logarithmic and its complexity is O(log n)

52
 Switch Execution
Switch: Take the complexity of the most expensive case
char key;
int[] X = new int[n];
int[][] Y = new int[n][n];
........
switch(key) {
case 'a':
for(int i = 0; i < X.length; i++) o(n)
sum += X[i];
break;
case 'b':
for(int i = 0; i < Y.length; j++) o(n2)
for(int j = 0; j < Y[0].length; j++)
sum += Y[i][j];
break;
} // End of switch block

Overall Complexity: O(n2)

53
Analyzing Method Calls

 The body of a loop may contain a call to a method

 To determine the order of the loop body, the order of

the method must be taken into account

 The overhead of the method call itself is generally

ignored

54
Analyzing Method Calls (cont…)
 Loop example:
 Suppose n is a multiple of 2. Determine the number of basic operations performed by of the
method myMethod():

static int myMethod(int n){ static int helper(int n){

int sum = 0; int sum = 0;
for(int i = 1; i < n; i = i * 2) for(int i = 1; i <= n;
sum = sum + i + helper(i); i++) sum = sum + i;
return sum; return sum;
} }
 Solution: The number of iterations of the loop:

for (int i = 1; i < n; i = i * 2)

sum = sum + i + helper(i);

is log2n

Hence the number of basic operations is:

1 + 1 + (1 + log2n) + log2n[2 + 4 + 1 + 1 + (n + 1) + n[2 + 2] + 1] + 1

= 3 + log2n + log2n[10 + 5n] + 1 = 5 n log2n + 11 log2n + 4

55
Analyzing Method Calls (cont…)

 Recursion:
 Analyze from the inside (or deepest part) first and work outwards. If there are
function calls, these must be analyzed first.

 Example:
long factorial (int n) Time Units to Compute:
{ if (n <= 1) return ------------------------
1; 1 for the test
1 for the multiplication statement
else return n * factorial (n – 1)
; } What about the function call?

 The running time of factorial(n) = T(n) = 2 + T(n-1) =

4 + T(n-2) = 6 + T(n-3) = … = 2n = O(n)

56
Examples of Algorithms and their big-O
complexity
Big-O Notation Examples of Algorithms

Push, Pop, Enqueue (if there is a tail reference), Dequeue,

O(1)
Accessing an array element

O(log(n)) Binary search

O(n) Linear search

O(n log(n)) Heap sort, Quick sort (average), Merge sort

O(n2) Selection sort, Insertion sort, Bubble sort

O(n3) Matrix multiplication

O(2n) Towers of Hanoi

57
58