Data Structures and Algorithms

(ESO207)

Lecture 5:
• More on Proof of correctness of an algorithm
• Design of O(𝑛) time algorithm for Local Minima in a grid

1
PROOF OF CORRECTNESS

2
What does correctness of an algorithm mean ?

For every possible valid input, the algorithm must output correct answer.

Let us take a toy algorithm J

3
 Algorithm for computing
sum of numbers from 𝟎 to 𝒏
Sum_of_Numbers(𝑛)
{ Sumß0;
for 𝑖 = 1 to 𝑛
{
Sumß Sum + 𝑖;
}
return Sum;
How will you convince any person
} that Sum_of_Numbers(𝑛)
indeed is correct ?
Natural responses:
• It is obvious !
• Compile it and run it for some random values of 𝑛.
• Go over first few iterations explaining what happens to Sum.
4
 How will you respond
if you have to do it for the following code ?

5
Think for some time to realize
• the non-triviality
• the Importance
of proof of correctness of an iterative algorithm.

In the following slide, we present an overview of

the proof of correctness.

Interestingly, such a proof will be just

Expressing our intuition/insight of the algorithm in a formal way J.

6
 Proof of correctness
For an iterative algorithm

Insight of the algorithm Theorem

Assertion Assertion
P(𝑖 − 1) P(𝑖)

Start 𝑖−1 𝑖 Iterations

1 2 3 4
of Loop

Prove P(𝑖) by What would you expect

at the end of 𝑖th
1. Assuming
? P(𝑖 − 1) iteration ?
2. Theorem
? Proof by induction
3. Body
? of the Loop
The most difficult/creative part of proof : To come?up with the right assertion P(𝑖)
7
 Algorithm for computing
sum of numbers from 𝟎 to 𝒏
{ Sumß0;
for 𝑖 = 1 to 𝑛
{
Sumß Sum + 𝑖;
}
return Sum;
}

Assertion P(𝑖) : At
? the end of 𝒊th iteration Sum stores the sum of numbers from 𝟎 to 𝒊.

Base case: P(0) holds.

Assuming P(𝑖 − 1), assertion P(𝑖) also holds.
P(𝑛) holds.
8
 An O(𝒏) time Algorithm for Max-sum subarray
Let S(𝑖): the sum of the maximum-sum subarray ending at index 𝑖.

Theorem 1 : If S(𝑖 − 1) > 0 then S(𝑖) = S(𝑖 − 1) + A[𝑖]

else S(𝑖) = A[𝑖]

Max-sum-subarray-algo(A[0 … 𝒏 − 𝟏])
{ S[0] ß A[0]
for 𝑖 = 1 to 𝒏 − 𝟏
{ If S[𝑖 − 1] > 0 then S[𝑖] ß S[𝑖 − 1] + A[𝑖]
else S[𝑖] ß A[𝑖]
}
“Scan S to return the maximum entry”
}
Assertion P(𝑖) : S[𝑖]
? stores the sum of maximum sum subarray ending at A[𝑖].
Homework: Prove that P(𝑖) holds for all 𝑖 ≤ 𝑛 − 1
9
LOCAL MINIMA IN A GRID

10
 Local minima in a grid

Definition: Given a 𝒏 × 𝒏 grid storing distinct numbers, an entry is local

minima if it is smaller than each of its neighbors.
𝒋

31
𝒊 5 3 10
99

11
 Local minima in a grid

Problem: Given a 𝒏 × 𝒏 grid storing distinct numbers, output any local

minima in O(𝒏) time. 𝒋

31
𝒊 5 3 10
99

12
 Two simple principles

1. Respect every new idea which solves a problem even partially.

2. Principle of simplification:
If you find a problem difficult,
è try to solve its simpler version, and then

è extend this solution to the original (difficult) version.

13
 A new approach
Repeat : if current entry is not local minima, explore the neighbor storing smaller
value.
j

i 3

14
 A new approach
Explore()
{ Let c be any entry to start with;
While(c is not a local minima)
{
c ß a neighbor of c storing smaller value
}
return c;
}

15
 A new approach
Explore()
{ Let c be any entry to start with;
While(c is not a local minima)
{
c ß a neighbor of c storing smaller value How to apply this
} principle ?
return c;
}
Worst case time complexity : O(𝑛! )

First principle: Second principle:

Do not discard Explore() Simplify the problem

16
 Local minima in an array
A local minima exists
in this region. Why ?
23
17

A 9 17 23
𝑖

Theorem: There is a local minima in A[0,…, 𝑖 − 1].

Proof: Suppose we execute Explore() from A[𝑖 − 1].
Explore(), if terminates, will return local minima.
It will terminate without ever entering A[𝑖,…, 𝑛 − 1].
Algorithmic proof
Hence there is a local minima in A[0,…, 𝑖 − 1].

17
 Local minima in an array

23
17

A 9 17 23
𝑖

Theorem: There is a local minima in A[0,…, 𝑖 − 1].

è We can confine our search for local minima to only A[0,…, 𝑖 − 1].
èOur problem size has reduced.
Question: Which 𝑖 should we select so as to reduce problem size significantly ?
Answer: 𝑚𝑖𝑑𝑑𝑙𝑒 point of array A.
18
 Local minima in an array
(Similar to binary search)

int Local-minima-in-array(A) { O(log 𝒏)

L ß 0;
How many
R ß 𝒏 − 𝟏; iterations ?
found ß FALSE;
while( not?? found )
{
𝒎𝒊𝒅 ß (L + R)/2;
If (𝒎𝒊𝒅 is a local minima)
O(𝟏) time
found ß TRUE; in one iteration
else if(A[𝒎𝒊𝒅 + 𝟏] < A[𝒎𝒊𝒅]) Lß 𝒎𝒊𝒅
?? + 𝟏 ;
else ??Rß 𝒎𝒊𝒅 − 𝟏
}
return 𝒎𝒊𝒅; }
Proof of correctness ?
è Running time of the algorithm = O(log 𝒏)
19
 Local minima in an array
(Proof of correctness)

What can you say about the

algorithm at the end of 𝑖th
iteration.
𝑳 𝑹

A
P(𝑖) : At the end of 𝑖th iteration,
“A local minima of array A exists in A[𝑳,…, 𝑹].”

20
 Local minima in an array
(Proof of correctness)

𝑳 𝑹

A
P(𝑖) : At the end of 𝑖th iteration,
“A local minima of array A exists in A[𝑳,…, 𝑹].”

=
“A[𝑳]< A[𝑳 − 𝟏]” and “A[𝑹]< A[𝑹 + 𝟏]”.
Homework:
• Make sincere attempts to prove the assertion P(𝑖).
• How will you use it to prove that Local-minima-in-array(A) outputs a local21
minima ?
 Local minima in an array
(Proof of correctness)

Theorem: A local minima in an array storing 𝒏 distinct elements

can be found in O(log 𝒏) time.

22
 Local minima in a grid
(extending the solution from 1-D to 2-D)
Under what
Search for a local minima in the column M[∗, 𝒎𝒊𝒅 ] A What
local minima
circumstances
exists
if thereeven
is nothis
inlocal
this minima
region. Why
in the?
smallest element is not
aentire
localcolumn.
minima ?
𝒎𝒊𝒅

Smallest element
Execute Explore()
of the column
from M[𝒊, 𝒎𝒊𝒅 + 𝟏 ]

𝒊 9 7

Homework:
Use this idea to design an O(𝒏 log 𝒏) time algorithm for this problem.
23
… and do not forget to prove its correctness J.
 Make sincere attempts to
answer all questions raised in this lecture.

24