Data Structures and Algorithms

(ESO207A)

Lecture 8
• Proof of Correctness.
• Design of O() time algorithm for Local Minima in a grid

1
RECAP FROM LAST LECTURE

2
 Proof of correctness
For an iterative algorithm

Insight of the algorithm Theorem

Assertion Assertion
P() P()

Start
of Loop 1 2 3 4 𝑖 −1 𝑖 Iterations

Prove P() by What would you expect

1. ?Assuming P() at the end of th
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()
3
 EXAMPLE
MAXIMUM SUM SUBARRAY

4
 An O() time Algorithm for Max-sum subarray

Max-sum-subarray-algo(A[]) Let S(): the sum of the maximum-sum

{ S[]  A[] subarray ending at index .
for = 1 to Theorem 1 :
{ If S[] > 0 then S[]  S[] + A[] If S() > 0 then S() = S() + A[]
else S[]  A[] else S() = A[]
}
“Scan S to return the maximum entry”
}

5
 An O() time Algorithm for Max-sum subarray
P() holds for each

Max-sum-subarray-algo(A[]) Let Opt(): the sum of the maximum-sum

{ S[]  A[] subarray ending at index .
for = 1 to Theorem 1 :
{ If S[] > 0 then S[]  S[] + A[] If Opt() > 0 then Opt() = Opt() + A[]
else S[]  A[] else Opt() = A[]
}
“Scan S to return the maximum entry”
}
P() : ? S[] = Opt()

Assume P() holds.

P() holds. S[] = Opt()
P() holds.
How to show that P() holds ?

6
 EXAMPLE
LOCAL MINIMA IN AN ARRAY

7
 Algorithm for Local minima in an array
P() holds for th iteration, if executed
int Local-minima-in-array(A) P() :
{ L; LThere
R is a local
A[L] <minima
A[L] amongA[R]
A[L],...,A[R].
< A[R]
R;
found  FALSE;
while( not ?? found )
{ mid  (L R)/;
If (mid is a local minima) found  TRUE;
else if(A[mid] < A[mid]) ?? mid;
L
else ??R mid1;
}
return mid;
} A

P()Assume
holds. P() holds.
L R L R

8
 Algorithm for Local minima in an array
P() holds for th iteration, if executed.
int Local-minima-in-array(A) P() :
{ L; L R A[L] < A[L] A[R] < A[R]
R;
found  FALSE;
while( not ?? found )
{ mid  (L R)/;
If (mid is a local minima) found  TRUE;
else if(A[mid] < A[mid]) ?? mid;
L
else ??R mid1;
}
return mid;
} A

Assume P() holds. P() holds.

L R L R

9
 Algorithm for Local minima in an array
P() holds for th iteration, if executed.
int Local-minima-in-array(A) P() :
{ L; L R A[L] < A[L] A[R] < A[R]
R;
found  FALSE;
while( not ?? found )
{ mid  (L R)/;
If (mid is a local minima) found  TRUE;
else if(A[mid] < A[mid]) ?? mid;
L
else ??R mid1;
}
return mid;
} A

Assume P() holds. P() holds.

L mid R
L R

10
 Algorithm for Local minima in an array
P() holds for th iteration, if executed.
int Local-minima-in-array(A) P() :
{ L; L R A[L] < A[L] A[R] < A[R]
R;
found  FALSE;
while( not ?? found )
{ mid  (L R)/;
If (mid is a local minima) found  TRUE;
else if(A[mid] < A[mid]) ?? mid;
L
else ??R mid1;
}
return mid;
} A

Assume P() holds.

L mid R
L R

11
 Algorithm for Local minima in an array
P() holds for th iteration, if executed.
int Local-minima-in-array(A) P() :
{ L; L R A[L] < A[L] A[R] < A[R]
R;
found  FALSE;
while( not ?? found )
{ mid  (L R)/;
If (mid is a local minima) found  TRUE;
else if(A[mid] < A[mid]) ?? mid;
L
else ??R mid1
}
return mid;
} A

Assume P() holds. P() holds.

mid L R
L R

12
 Algorithm for Local minima in an array
P() holds for th iteration, if executed.
int Local-minima-in-array(A) P() :
{ L; L R A[L] < A[L] A[R] < A[R]
R;
found  FALSE;
while( not ?? found )
{ mid  (L R)/;
If (mid is a local minima) found  TRUE;
else if(A[mid] < A[mid]) ?? mid;
L
else ??R mid1
}
return mid;
} A

Assume P() holds.

L mid R
L R

13
 Algorithm for Local minima in an array
P() holds for th iteration, if executed
int Local-minima-in-array(A) P() :
{ L; L R A[L] < A[L] A[R] < A[R]
R;
found  FALSE;
while( not ?? found )
{ mid  (L R)/;
If (mid is a local minima) found  TRUE;
else if(A[mid] < A[mid]) ?? mid;
L
else ??R mid1
}
return mid;
} A

Assume P() holds. P() holds.

L R L R mid

14
 Algorithm for Local minima in an array
P() holds
P()for
holds
th iteration,
for each if executed
int Local-minima-in-array(A) P() :
{ L; L R A[L] < A[L] A[R] < A[R]
R;
found  FALSE;
while( not ?? found ) Question 1:
{ mid  (L R)/; What happens during an iteration
If (mid is a local minima) found  TRUE; that begins with L R.
else if(A[mid] < A[mid]) ?? mid; Answer:
L
else ??R mid1 Local Minima is found 
}
return mid;
}

15
 Algorithm for Local minima in an array
P() holds for th iteration, if executed
int Local-minima-in-array(A) P() :
{ L; L R A[L] < A[L] A[R] < A[R]
R;
found  FALSE;
while( not ?? found ) Question 2:
{ mid  (L R)/; During an iteration that begins with L R,
If (mid is a local minima) found  TRUE; if local minima is not found,
else if(A[mid] < A[mid]) ?? mid; what happens to R L ?
L
else ??R mid1 Answer:
} Strictly decreases.
return mid;
}

Algorithm is correct.

16
 Local minima in an array
(Proof of correctness)

Theorem: A local minima in an array storing distinct elements

can be found in O(log ) time.

17
Local minima in a GRID
in O(log ) time

18
 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[, ] AWhat
local minima
circumstances
exists
if thereeven
is no this
inlocal
this minima
region. Why
in the?
smallest element is not
aentire
localcolumn.
minima ?
Explore() from M[, ] 𝒎𝒊𝒅

Smallest element
of the column

𝒊 9 7

19
 Local minima in a grid
Int Local-minima-in-grid(M) // returns the column containing a local minima
{ L  0;
R;
found  FALSE;
while(not found)
{  (L + R)/2;
If (M[*, ] has a local minima) found  TRUE; O() time
else {
let M[, ] be the smallest element in M[*, ] O() time
O()
if(M[, ] < M[, ]) ?? ; L
time
else ?? R
}
}
return ;
} Proof of correctness ?
 Running time of the algorithm = O(log )
Homework 20
 Local minima in a grid

Theorem: A local minima in an × grid storing distinct elements

can be found in O( log ) time.

How to improve it to O()

?

21
 Local minima in a grid in O() time
Let us carefully look at the calculations of the running time of the current algo.

= O() (log terms)

What about the following series

=? (log terms)

It is = O() .

Get an !DEA from this series to modify our current algorithm

22
 Local minima in a grid in O() time

Bisect alternatively along rows and column

Try on your own.

We shall revisit it
towards the end of
the course.

23
 Quiz 1
• Date: 20th August, 2022
• Time: 9:00 AM to 10:45 AM

Details of seating plan will be posted today.

24