Analysis of algorithms

CS 1401
ANALYSIS OF ALGORITHMS (CS 1401)

UNIT 1:
Introduction, Review of basic concepts,
Advanced data structures like Binomial
Heaps, Fibonacci Heaps 5(L)
UNIT 2:
Divide and Conquer with examples such as
Sorting, Matrix
Multiplication, Convex hull etc 6(L)
UNIT 3:
Dynamic programming with examples such
as Kanpsack, All pair shortest paths etc
4(L)

UNIT 4:
Backtracking, Branch and Bound
with examples such as Travelling Salesman
Problem etc 6(L)
UNIT 5:
Algorithms involving Computational
Geometry 4(L)

UNIT 6:
Selected topics such as NP completeness,
Approximation algorithms, Randomized
algorithms, String Matching 5(L)
Text/Reference Books

Introduction to Algorithms by
Thomas H. Coreman, Charles E. Leiserson
and Ronald L. Rivest

Fundamentals of Computer Algorithms by

E. Horowitz & S Sahni

The Design and Analysis of Computer

Algorithms by Aho, Hopcraft, Ullman
ALGORITHM

A finite set of instructions which if followed

accomplish a particular task.
In addition every algorithm must satisfy
following criteria:
1. Input: zero or more quantities externally
supplied
2. Output: at least one quantity is produced
3. Definiteness: Each instruction must be
clear and unambiguous.
4. Finiteness: In all cases algorithm must
terminate after finite number
of steps.
5. Effectiveness: each instruction must be
sufficiently basic.
• Two algorithms on two systems
• Algorithm A1 50 n lg n
• Algorithm A2 2 n2

A2 Super computer
108 ins/sec

A1
P.C
106 ins /sec
For n = 106

Time taken by Super Computer

= 2.(106)2/ 108 = 20,000 sec

Time taken by P .C.

= 50 . 106 lg 106 / 106 = 1,000 sec
Thus by using a fast algorithm , the personal
computer gives results
20 times faster than the result given by super
computer using a slow algorithm.

Thus a good algorithm is like a sharp knife, it

does exactly what it is supposed to do
with a minimum amount of effort.
Complexity
Some questions to answer:
 How fast can we solve a
problem?
 There may be many algorithms for a
given problem. Which algorithm to use?
 What are the classical algorithm
design techniques ?
 Are there problems inherently difficult
to solve?
How do we express the complexity of algorithm?

Resources : Time and Space

Complexity lower bounds for problems.

Complexity classes P, NP etc.

Pseudocode

• Pseudocode is an English language like

representation of the code required for an algorithm.

• It is partly English, partly structured code.

• The English part provides a relaxed syntax that is

easy to read.

• The code part consists of an extended version of the

basic algorithmic constructs-sequence, selection
and iteration.
Sequence, selection, loop
• A sequence is a series of statements that do not
alter the execution path within an algorithm.
• Statements such as assign and add are sequence
statements.
• A call to another algorithm is also considered a
sequence statement.
• Selection statements evaluate one or more
alternatives. Paths are followed based on its
result.
• The typical selection statement is the two
way selection
• if (condition) action 1 else action 2.
• The part of the loop are identified by
indentation.
• Loop iterates a block of code. It closely
resembles the while loop. It is a pretest loop.
Example
Algorithm deviation
It finds deviations from average.

Pre: nothing
Post: numbers are read and deviation
from average printed
1 i= 0
2 Loop(all data are read)
1 I=I+1
2 read numbers into array[i]
3 sum = sum + number
3 Average = sum / I
4 Print (average)
5 J=0
6 Loop (j < i)
1 j = j+ 1
2 dev = array[j] – average
3 print (array [ j] . Dev)
7 Return
8 End deviation
Analysis vs. Design

• Analysis: predict the cost of an algorithm in

terms of resources and performance

• Design: design algorithms which minimize

the cost
Time Complexity
Real Time:

To analyze the real time complexity of a program

we need to determine two numbers for each
statement in it:

• amount of time a single statement will take.

• No. of times it is executed.

• Product of these two, will be the total time taken

by the statement.
First no. depends upon the machine and
compiler used , hence the real time
complexity is machine dependent.
Machine model: Generic Random Access
Machine (RAM)

• Executes operations sequentially

• Set of primitive operations:
• Arithmetic. Logical, Comparisons, Function calls

• Simplifying assumption: all ops cost 1 unit

• Eliminates dependence on the speed of our computer,
otherwise impossible to verify and to compare
Frequency count

• To make analysis machine independent

it is assumed that every instruction takes
the same constant amount of time for
execution.

• Hence the determination of time

complexity of a program is the matter of
summing the frequency counts of all the
statements.
Linear loop
1 i=1
2 Loop(I <= 1000)

1 application code
2 I=I + 1

The body of the loop is repeated 1000 times.

1 I=1
2 Loop (I <= n)
1 Application code
2 I=I+2

For this the code time is proportional to n

Logarithm Loops
Multiply loops Divide loops
1 I=1 1 I=n
2 Loop (I < n) 2 loop( i>= 1)
1 application code 1 application
code 2 I=I/2
2 I = i*2

F(n) = [log n] F(n) = [log n]

Nested loop- linear logarithmic

• 1 I=1
• 2 loop(I <= n)
1 j=1
2 loop( j < = n)
1 application code
2 j = j *2
3 I=I+1

F(n ) = [n log n]
Dependent Quadratic
1 I=1
2 loop ( I < = n)
1 j=1
2 loop( j < = i)
1 application code
2 j=j+1
3 I=I + 1
no of iterations in the body of the inner loop is
1 + 2 + 3 + 4 +… + n I.e. n(n +1)/2

On an average = ( n+1/)2
thus total no of iterations = n (n+1)/2
Quadratic
1 i=1
2 Loop (I < = n)
1 j=1
2 Loop( j < = n)
1 application code
2 j = j+1
3 I = i+1

F(n) = n2
Insertion sort

INSERTION-SORT (A, n) ⊳ A[1 . . n]

for j ← 2 to n
do key ← A[j]
i←j–1
“pseudocode” while i > 0 and A[i] > key
do A[i+1] ← A[i]
i←i–1
A[i+1] ← key
1 j i n
A:
key
sorted
Insertion Sort
// At start, the singleton sequence A[1] is trivially
sorted
Freq.
1 for j  2 to length[A] n
2 do key  A[j] n-1
* insert A[j] into sorted array
* sequence A[1….j-1]
3 I  j-1 n-1
// let tj be the number of times the following
//while-loop is tested for the value j
4 while I> 0 and A[I] > key tj
5 Do A[I+1]  A[I] (tj-1)
6 I  I-1 (tj–1)
7 A[I+1]  key n-1
T(n) =

c1 n + c2 (n-1) + c3(n-1)+

c4 tj + c5 (tj-1) +

c6 (tj-1) + c7 (n-1)
• If Cj = 1

• T(n) = n + (n-1) + (n-1) + tj + 2 (tj-1) + (n-1)

• = n + 3(n-1) + 3tj -2(n-1)
• T(n) = 2n + 3  (tj ) for j = 1 to n-1

• = O( n2 )
• average case:
xj is being inserted into the sorted sequence
x1 x2 ... x j-1
• the probability that xj is the largest: 1/j
• takes 2 data movements
• the probability that xj is the second largest : 1/j
• takes 3 data movements
:
• # of movements for inserting xj: 1 4 7 5

2 3 j 1 j  3
  
j j j 2
n
j  3 ( n  8)( n  1)
M    O(n 2 )
j 2 2 4
Worst Case and best case
Best case: file is already sorted
in this case tj = 1

so total time is a linear function of n

Worst case: if file is in the reverse order

in this case tj = j-1

so total time is a quadratic function of n

Binary search

• sorted sequence : (search 9)

1 4 5 7 9 10 12 15
step 1 
step 2 
step 3 
• best case: 1 step = O(1)
• worst case: (log2 n+1) steps = O(log n)
• average case: O(log n) steps
n cases for successful search
n+1 cases for unsuccessful search

Average # of comparisons done in the binary tree:

1  k i 1 
A(n) = 2 n  1   i 2  k ( n  1), where k = log n+1
i 1
k
Assume
k
n=2 proved by induction
 i 2 i 1  2 k ( k  1)  1 on k
i 1

1
A(n) = (( k  1) 2 k  1  k(2 k  1))
2n  1
k as n is very large
= log n
= O(log n)
Quick sort

11 5 24 2 31 7 8 26 10 15
↑ ↑
11 5 10 2 31 7 8 26 24 15
↑ ↑
11 5 10 2 8 7 31 26 24 15
△• Recursively apply the
△ same procedure.
7 5 10 2 8 11 31 26 24 15
|← <11 →| |← > 11 →|
Best case : O(n log n)

A list is split into two sublists with almost equal size.

log n rounds are needed.

In each round, n comparisons (ignoring the element used to
split) are required.
Worst case : O(n2)

In each round, the number used to split is either the

smallest or the largest.
n(n  1)
n  (n  1)    1   O(n 2 )
2
Average case: O(n log n)
s n-s
include the splitter
T(n) = Avg (T(s)  T( n  s))  cn
1 s n
1 n
= 
n s 1
( T ( s )  T(n  s) )  cn

1
= n
(T(1)+T(n1)+T(2)+T(n2)+…+T(n)+T(0))+cn, T(0)=0
1
= n
(2T(1)+2T(2)+…+2T(n1)+T(n))+cn
(n1)T(n) = 2T(1)+2T(2)+…+2T(n1) + cn2……(1)
(n2)T(n-1)=2T(1)+2T(2)+…+2T(n2)+c(n1)2…(2)

(1)  (2)

(n1)T(n) (n2)T(n1) = 2T(n1)+c(2n1)

(n1)T(n) nT(n1) = c(2n1)

T(n) T(n  1) 1 1
=  c (  )
n n 1 n n 1
1 1 1 1 1
=c( n  n  1 )+c( n  1  n  2 )+…+c( 2  1 )+T(1), T(1) = 0
1 1 1 1 1
=c( n n  1 2 )+c( n  1 n  2 ...1)
  ... 
• <= 2log(n)+ constant

• Thus
• T(n) <= n log(n)
Example of insertion sort

8 2 4 9 3 6
2 8 4 9 3 6
2 4 8 9 3 6
2 4 8 9 3 6
2 3 4 8 9 6
2 3 4 6 8 9 done
Worst/ best/average cases

• Worst case is the longest running time for

any input of size n
• O-notation represents upper bound I.e. an
upper bound for worst case.
• Best case is time taken for some input
data set that results in best possible
performance.You cannot do better. This is
lower bound.
• Average case is the average performance
Asymptotic Notation
• Goal: to simplify analysis by getting rid of
unneeded information (like “rounding”
1,000,001≈1,000,000)
• We want to say in a formal way 3n2 ≈ n2
•The “Big-Oh” Notation:

•given functions f(n) and g(n), we

say that f(n) is O(g(n) ) if and only
if there are positive constants c
and n0 such that f(n)≤ c g(n) for
n ≥ n0
f(n) is O(g(n) if

f ( n)
lim n is finite.
g ( n)
Example
For functions f(n) f(n) = 2n + 6
and g(n) (to the
right) there are
positive constants c
and n0 such that:
f(n) ≤ c g(n)
for n ≥ n0
conclusion:
2n+6 is O(n).
Another Example

On the other hand…

n2 is not O(n) because
there is no c and n0 such
that: n2 ≤ c n for n ≥ n0

(As the graph to the right

illustrates, no matter how
large a c is chosen there is
an n big enough that
N2 > c n ) .
Asymptotic Notation (cont.)

• Note: Even though it is correct to say “7n - 3 is O(n3)”, a

better statement is “7n - 3 is O(n)”, that is, one should
make the approximation as tight as possible

• Simple Rule: Drop lower order terms and constant

factors
7n-3 is O(n)
8n2log n + 5n2 + n is O(n2log n)
Asymptotic Notation (terminology)
• Special classes of algorithms:

constant O(1)
logarithmic: O(log n)
linear: O(n)
quadratic: O(n2)
polynomial: O(nk), k ≥ 1
exponential: O(an), a > 1
EXAMPLE

Consider 1/3 n2 – 5n

The dominating term is n2

Therefore it should be of O(n2)

•Given a positive constant c , a positive integer

n0 to be found such that
1/3 n2 - 5 n  c n2
Dividing the inequality throughout by n2

1/3 - 5/n  c

Therefore if n0  1 , choosing c  1/3 the

inequality will never be violated

Hence the expression is indeed of O(n2)

Polynomial of degree k is O( nk)
T(n) = ak nk + ak-1 nk-1 +……..a0 is O( nk)
T(n) <= |ak| nk +| ak-1| nk-1 +…….|a0|
<= nk(|ak| n0 +| ak-1| n-1 +…….|a0|n-k)
<= nk(|ak| n0 +| ak-1| n-1 +…….|a0|n-k) for n >1
Since n >1 n I-k <1 for I 0 <= I <= k
Therefore
T(n) <= nk ( |a0| + |a1 | + …. |ak |)
let c = ( |a0| + |a1 | + …. |ak |)
And f(n) = nk
T(n) <= c f(n) for n >1
Thus T(n) = O( nk)
•“Relatives” of the Big-Oh

• (f(n)): Big Omega--asymptotic

lower bound
• (f(n)): Big Theta--asymptotic tight
bound
Little o
• Informally, saying some equation f(n) = o(g(n))
means f(n) becomes insignificant relative to g(n)
as n approaches infinity. The notation is read, "f of
n is little oh of g of n".

• Formal Definition: f(n) = o(g(n)) means for all c > 0

there exists some k > 0 such that 0 ≤ f(n) < cg(n)
for all n ≥ k. The value of k must not depend on n,
but may depend on c
f ( n)
lim n 0
g ( n)

3n+2 =o(n2 )

3n  2
since lim n 2
0
n

3n+2 ≠ o(n) while 3n+2 =O(n)

Asymptotic Analysis of The Running Time
• Use the Big-Oh notation to express the
number of primitive operations executed
as a function of the input size.
• Comparing the asymptotic running time
-an algorithm that runs in O(n) time is
better than one that runs in O(n2) time
-similarly, O(log n) is better than O(n)
-hierarchy of functions: log n << n << n2
<< n3 << 2n
Caution! Beware of very large constant factors.

An algorithm running in time 1,000,000 n is still

O(n) but might be less efficient on your data set

than one running in time 2n2, which is O(n2)

Properties of the O notation
• Constant factors may be ignored
 " k > 0, kf is O( f)
Properties of the O notation
• Constant factors may be ignored
 " k > 0, kf is O( f)
• Higher powers grow faster
• nr is O( ns) if 0  r  s
Properties of the O notation
• Constant factors may be ignored
 " k > 0, kf is O( f)
• Higher powers grow faster
• nr is O( ns) if 0  r  s
Fastest growing term dominates a sum
• If f is O(g), then f + g is O(g)
eg an4 + bn3 is O(n4 )
Properties of the O notation
• Constant factors may be ignored
 " k > 0, kf is O( f)
• Higher powers grow faster
• nr is O( ns) if 0  r  s
. Fastest growing term dominates a sum
• If f is O(g), then f + g is O(g)
eg an4 + bn3 is O(n4 )
. Polynomial’s growth rate is determined by leading
term
• If f is a polynomial of degree d,
then f is O(nd)
Properties of the O notation
• f is O(g) is transitive
• If f is O(g) and g is O(h) then f is O(h)
Properties of the O notation

• Product of upper bounds is upper bound for the

product
• If f is O(g) and h is O(r) then fh is O(gr)
Properties of the O notation
• f is O(g) is transitive
• If f is O(g) and g is O(h) then f is O(h)
• Product of upper bounds is upper bound for the
product
• If f is O(g) and h is O(r) then fh is O(gr)
• Exponential functions grow faster than powers
• nk is O( bn ) " b > 1 and k  0
eg n20 is O( 1.05n)
Properties of the O notation
• f is O(g) is transitive
• If f is O(g) and g is O(h) then f is O(h)
• Product of upper bounds is upper bound for the
product
• If f is O(g) and h is O(r) then fh is O(gr)
• Exponential functions grow faster than powers
• nk is O( bn ) " b > 1 and k  0
eg n20 is O( 1.05n)
• Logarithms grow more slowly than powers
• logbn is O( nk) " b > 1 and k > 0
eg log2n is O( n0.5)
• Thus {log2(n)}2 is O(nk) for any k > 0
Properties of the O notation
• f is O(g) is transitive
• If f is O(g) and g is O(h) then f is O(h)
• Product of upper bounds is upper bound for the
product
• If f is O(g) and h is O(r) then fh is O(gr)
• Exponential functions grow faster than powers
• nk is O( bn ) " b > 1 and k  0
eg n20 is O( 1.05n)
• Logarithms grow more slowly than powers
• logbn is O( nk) " b > 1 and k > 0
eg log2n is O( n0.5)
Important!
Properties of the O notation
• All logarithms grow at the same rate
• logbn is O(logdn) " b, d > 1
Properties of the O notation
• All logarithms grow at the same rate
• logbn is O(logdn) " b, d > 1

• Sum of first n rth powers grows as the (r+1)th power

n
  kr is  ( nr+1 )
k=1
n
eg  i = n(n+1) is  ( n2 )
k=1 2
Polynomial and Intractable Algorithms
• Polynomial Time complexity
• An algorithm is said to be polynomial if it is
O( nd ) for some integer d
• Polynomial algorithms are said to be efficient
• They solve problems in reasonable times!

• Intractable algorithms
• Algorithms for which there is no known
polynomial time algorithm
We will come back to this important class later in
the course
Categories of algorithm efficiency
Efficiency Big O
Constant O(1)
Logarithmic O(log n)
Linear O(n)
Linear logarithmic O(n log n)
Quadratic O(n2)
Polynomial O(nk)
Exponential O(cn)
Factorial O(n!)
-notation
For function g(n), (g(n)) is
given by:

(g(n)) = {f(n):  +ve

constants c1, c2, and n0 such
that 0  c1g(n)  f(n) 
c2g(n),
"n  n0 }

Intuitively: Set of all functions that

have the same rate of growth as g(n).

g(n) is an asymptotically tight bound for f(n).

O -notation
For function g(n), O(g(n)) is
given by:

O(g(n)) = {f(n):  +ve

constants c and n0 such that 0 
f(n)  cg(n), "n  n0 }

Intuitively: Set of all functions

whose rate of growth is the same as
or lower than that of g(n).
g(n) is an asymptotic upper bound for f(n).
f(n) = (g(n))  f(n) = O(g(n)).
(g(n))  O(g(n)).
 -notation
For function g(n), (g(n)) is
given by:
(g(n)) = {f(n):  +ve
constants c and n0 such that 0 
cg(n)  f(n), "n  n0 }
Intuitively: Set of all
functions whose rate of
growth is the same as or
higher than that of g(n).
g(n) is an asymptotic lower bound for f(n).
f(n) = (g(n))  f(n) = (g(n)).
(g(n))  (g(n)).
Relations Between , O, 
Practical Complexity

250

f(n) = n
f(n) = log(n)
f(n) = n log(n)
f(n) = n^2
f(n) = n^3
f(n) = 2^n

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Problem size
10 102 103 104

log2n 3.3 6.6 10 13.3

n 10 102 103 104

nlog2n 0.33x10 0.7x103 104 1.3x105

n2 102 104 106 108

2n 1024 1.3x103 >10100 >10100

n! 3x106 >10100 >10100 >10100

• Q1 Find functions f and g such that
• (i) f is O (g)
• (ii) f is not Θ (g)
• (iii) f(n) > g(n) for infinitely many n.

• Let us take g(n) = ½ for all n

• f(n) be a function with values in(0,1) close
to 1.
• f(n) = sin(n) or
• f(n) = 1/n if n is perfect square and 1
otherwise.
Exercise 1
If the complexity of the algorithm DOIT can
be expressed as O(n2)
Calculate the complexity of the following
program segment:
I=1
loop ( I < n)
DOIT(…)
I = I* 2
Answer : O(n2 log2n)
Exercise
Reorder the following complexity from
smallest to largest:
(a) N log2(n)
(b) n + n2 + n3
(c) 24
(d) n
Answer :
(a) 24
(b) n
(c) N log2(n)
(d) n + n2 + n3
Find Two non negative functions f(n)
and g(n) such that neither f(n) = O(g(n))
nor g(n) is O(f(n))
Let f(n) = 1+ cos (n)
g(n) = 1+ sin(n)

when f(n) = 0, g(n) = 2

And
when g(n) = 0 , f(n) = 2
it is not possible to find a positive constant c such that f(n) <=
g(n) for large n . Because g(n) always comes back to zero when
f(n) is 2
Is 2n+1 = O(2n )? Is 2 2n = O( 2 n)?
• Is 2n+1 = O(2n )? Yes!
• For any c >= 2

• 2n+1 <= C(2n ) for all n >= 1

• However

• Is 2 2n = O( 2 n)? Not correct.

• Lim 2 2n / ( 2 n ) is not finite.