6 Efficiency and

C
Computational
i
lC
Complexity
l i
COL 100 - Introduction to Computer
p
Science
II Semester 2015-2016
Department of Computer Science and Engineering
Indian Institute of Technology Delhi

What is a good algorithm/program?

Solution is simple but powerful/general

Easily understood by reader
Easily modifiable and maintainable
Correct for clearly defined situations
Efficient in space and time
Well documented
usable by those who do NOT understand the detailed
working

Portable
P t bl across computers
t
Can be used as a sub-program
Jan 2016

P. R. Panda, IIT Delhi

Efficiency of Algorithms
Algorithms/Programs evaluated in terms
of:
execution time
memory space

Jan 2016

P. R. Panda, IIT Delhi

Identifying Redundant
Computation
Redundant
computation
can be moved
from loop
loop invariant
loop-invariant
computation

Jan 2016

x = 0;
0
for (i = 0; i < 10; i++) {
y = (a*a*a + c)*x*x + (b*b)*x + c;
cout << y;
x = x + 0.01;
}

x = 0;
t1 = (a*a*a + c);
t2 = b*b;
for (i = 0; i < 10; i++) {
y = t1*x*x
1* * + t2*x
2* + c;
cout << y;
x = x + 0.01;
}
P. R. Panda, IIT Delhi

Early
l Termination
Searching in array:
break when found

Jan 2016

P. R. Panda, IIT Delhi

Computational Complexity
Quantitative measure of algorithms performance
needed
independent of programming language
independent of machine

Performance is characterised in terms of size of the

problem being solved
if problem size is n (e
(e.g.,
g searching in array of n integers)
how many operations are performed by algorithm?
as a function of n
indirectly measures execution time

how much memory is required for storage

as a function of n
Jan 2016

P. R. Panda, IIT Delhi

Rate of growth of functions

y = 2x

y = ax2 + b
y = ax + b
y = log x

Jan 2016

P. R. Panda, IIT Delhi

Asymptotic Analysis
What happens for large n?

Jan 2016

P. R. Panda, IIT Delhi

Order Notation
Function g(n) is of order O(f(n)) if
there exists c for which g(n) cf(n)
for all n some n1

Jan 2016

P. R. Panda, IIT Delhi

Rate of Growth of Functions

Given f(n) and g(n), which grows faster?
If Limnf(n)/g(n) = 0, then g(n) is faster
If Limnf(n)/g(n) = , then f(n) is faster
If Limnf(n)/g(n) = non-zero constant, then both
grow at the same rate

Two polynomials of the same degree grow at

the same rate
O(1) means constant time
independent
p
of n
Jan 2016

P. R. Panda, IIT Delhi

10

Why use O(
O( ) for measuring
Complexity?
p
y
Hiding constants
crude/approximate
easier to compute
holds across machines

How does algorithm scale for increasing n?

Which algorithm is better for large problem
s e
size?

Jan 2016

P. R. Panda, IIT Delhi

11

Computing the Complexity

Estimate the number of operations
as a function of input size

One for initialisation

Loop
oop executes
e ecu es n times
es
Max. of one comparison and one
assignment in each iteration
Max 2n operations for loop
Max.

int main () {
...
max = -10000;
for (i = 0; i < n; i++)
if (max < a[i])
max = a[i];
}

Total operations = 1 + 2n
Worst Case Complexity is O(n)
Best Case?
Average
g Case?

Jan 2016

P. R. Panda, IIT Delhi

12

Efficient Algorithms
Problem:
Given real number x and integer n
Write an algorithm to calculate xn

Jan 2016

P. R. Panda, IIT Delhi

13

First Algorithm
Power
power(x,n) = 1 for n = 0
power(x, n) = x * power(x,n-1) for n>1
float prod = 1.0;
for ((int i = 0;; i < n;; i++))
prod = prod * x;

Jan 2016

P. R. Panda, IIT Delhi

14

Efficient Algorithm
Fast Power
fpower(x,n) = 1 for n = 0
fpower(x,n) = x * (fpower(x, n/2))2 if n is odd
fpower(x,n) = (fpower(x, n/2))2 if n is even

What is the
complexity
p
y
of this algorithm?

Jan 2016

P. R. Panda, IIT Delhi

15

Searching in an Array
Given array A[N] and value V, is V
present in the Array? 0 1 2 3
A

fo nd = false;
found
for (i = 0; i < N; i++)
if (A[i] == V) {
found = true; // V is present
break;
}
Jan 2016

H many comparisons?
i
?
How
best case
average case
worst case

P. R. Panda, IIT Delhi

16