�ݺ�ߣ

Asymptotic Analysis of
Algorithms
SHYAMAL KEJRIWAL

Asymptotic Analysis of Algorithms
An algorithm is any well-de?ned step-by-step
procedure for solving a computational problem.
Examples -
Problem Input Output
Checking if a number is prime A number Yes/No
Finding a shortest path between your hostel and
your department
IITG Map, your hostel
name, your dept name
Well-defined shortest
path
Searching an element in an array of numbers An array of numbers Array index

An algorithm can be speci?ed in English, as a computer
program, or even as a hardware design.
There can be many correct algorithms to solve a given
problem. Is the correctness of an algorithm enough for
its applicability? By the way, how do we know if an
algorithm is correct in the first place?

The correctness of an algorithm is not enough. We
need to analyze the algorithm for efficiency.
Efficiency of an algorithm is measured in terms of:
?Execution time (Time complexity) �C How long does the
algorithm take to produce the output?
?The amount of memory required (Space complexity)

Let us analyze a few algorithms for space and time
requirements.
Problem: Given a number n (<=1) as an input, find the
count of all numbers between 1 and n (both inclusive)
whose factorials are divisible by 5.

ALGORITHM BY ALIA
1. Initialize current = 1 and count = 0.
2. Calculate fact = 1*2*��*current.
3. Check if fact%5 == 0. If yes, increment count
by 1.
4. Increment current by 1.
5. If current == n+1, stop. Else, go to step 2.
ALGORITHM BY BOB
1. Initialize current = 1, fact = 1 and count = 0.
2. Calculate fact = fact*current.
by 1.

ALGORITHM BY ALIA
by 1.
RUNNING TIME ANALYSIS
First note that there are few basic operations
that are used in this algorithm:
1. Addition (+)
2. Product (*)
3. Modulus (%)
4. Comparison (==)
5. Assignment (=)
We assume that each such operation takes
constant time ��c�� to execute.

ALGORITHM BY ALIA
by 1.
Note that different basic operations don��t take
same time for execution. Also, the time taken by
any single operation also varies from one
computation system to other(processor to
processor). For simplicity, we assume that each
basic operation takes some constant time
independent of its nature and computation system
on which it is executed.
After making the above simplified assumption to
analyze running time, we need to calculate the
number of times the basic operations are
executed in this algorithm. Can you count them?

ALGORITHM BY ALIA
by 1.

ALGORITHM BY BOB
by 1.
Now find the time complexity of the
other algorithm for the same problem
similarly and appreciate the difference.

ALGORITHM BY BOB
by 1.

CAN WE DO BETTER
Verify if the following algorithm is correct.
If n < 5, count = 0. Else, count = n - 4.
A constant time algorithm, isn��t it? For
such algorithms, we say that the time
complexity is ��(1).
Algorithms whose solutions are
independent of the size of the
problem��s inputs are said to have
constant time complexity. Constant
time complexity is denoted as ��(1).

The previous examples suggest that we need to know the
following two things to calculate running time for an
algorithm.
1. What are the basic or constant time operations present
in the computation system used to implement the
algorithm?
2. How many such basic operations are performed in the
algorithm? We count the number of basic operations
required in terms of input values or input size.

Similarly, to calculate space requirements for an algorithm,
we need to answer the following.
1. What are the basic or constant space data types present
in the computation system used to implement the
algorithm?
2. How many instances of such basic data types are used in
the algorithm?

Our computer systems generally support following constant
time instructions:
1. arithmetic (such as add, subtract, multiply, divide,
remainder, ?oor, ceiling)
2. data movement (load, store, copy)
3. control (conditional and unconditional branch,
subroutine call and return).

The basic data types generally supported in our computer
systems are integer and ?oating point (for storing real
numbers). Note that characters are also stored as 1 byte
integers in our systems.
So, if we use few arrays of ��n�� integers in our algorithm, we
can say that the space requirements of the algorithm is
proportional to n. Or, the space complexity is ��(n).

int x = 0;
for ( int j = 1; j <= n/2; j++ ){
x = x + j;
}
int n;
int *array;
cin >> n;
array = new int [n];
for ( int j = 0; j < n; j++ ){
array[j] = j + 1;
}
Now that you know the basic operations and data types present in modern
systems, can you analyze the running time and space requirements of the
following C++ code snippets.

�� Notation
To determine the time complexity of an algorithm:
Express the amount of work done as a sum f1(n) + f2(n) + �� + fk(n)
Identify the dominant term: the fi such that fj is ��(fi) and for k different from j
fk(n) < fj(n) (for all sufficiently large n)
Then the time complexity is ��(fi)

Let us try analyze a few more C++ codes and express the time and
space complexities in �� notation.
With independent nested loops: The number of iterations of the
inner loop is independent of the number of iterations of the outer
loop
int x = 0;
for ( int j = 1; j <= n/2; j++ ){
for ( int k = 1; k <= n*n; k++ ){
x = x + j + k;
}
}

Let us try analyze a few more C++ codes and express the time
and space complexities in �� notation.
With dependent nested loops: Number of iterations of the
inner loop depends on a value from the outer loop
int x = 0;
for ( int j = 1; j <= n; j++ ){
for ( int k = 1; k <= 3*j; k++ ){
x = x + j;
}
}

? C++ codes for algorithms discussed and running time analysis
? Linear search vs binary search
? Worst case time complexity

�ݺ�ߣ

Asymptotic analysis of algorithms.pptx

More Related Content

Asymptotic analysis of algorithms.pptx