Time Complexity of Recursive Algorithm – Divide and Conquer

Recurrence Relation
Dividing Functions
Time Complexity of Recursive Algorithm Example 1

Consider the following algorithm

Recurrence Relation:
BinarySearch ( A, start, end, x) {
T(n) = T(n/2) +
if (start < end) {
mid = (start +end)/2
if (x = = A[mid])
return mid;
else if (x > A[mid]) n/2
return BinarySearch ( A, mid +1, end , x)
else or
return BinarySearch ( A, start, mid -1 , x)
}
Single call at each activation
}
Time Complexity of Recursive Algorithm Example 1

Consider the following algorithm

Recurrence Relation: for n>1
BinarySearch ( A, start, end, x) {
T(n) = T(n/2) + 1
if (start < end) {
mid = (start +end)/2
if (x = = A[mid]) Constant
return mid; Only Dividing Cost
else if (x > A[mid]) No Combing here.
return BinarySearch ( A, mid +1, end , x)
else
return BinarySearch ( A, start, mid -1 , x)
}
} f(n) denotes cost of the work done outside the recursive call, which includes:
 cost of dividing the problem
 cost of combining/merging the solutions
Time Complexity of Recursive Algorithm Example 1

Consider the following algorithm

Recurrence Relation: for n>1
BinarySearch ( A, start, end, x) {
T(n) = T(n/2) + 1
if (start < end) {
mid = (start +end)/2 T(n) = 1 for n=1
if (x = = A[mid])
return mid;
else if (x > A[mid])
return BinarySearch ( A, mid +1, end , x)
else
return BinarySearch ( A, start, mid -1 , x)
}
}
Time Complexity of Recursive Algorithm Example 1

T(1) = 1 for n=1 𝑛

T(n)¿ 𝑇 ( 2 )+ 1
𝑛
T(n)¿ 𝑇 ( )
2 +1
( )
1 𝑛
¿𝑇 +1
22
] +1
𝑛
¿𝑇 ( 𝑛
22 )+ 2 2 T(n)¿ 𝑇 (
2
)+ 1

] 𝑛
2 ¿𝑇 ( )
23 +
¿𝑇 ( +3 ) 𝑛
23 3

¿𝑇 (
2𝑘)
𝑛
k
Time Complexity of Recursive Algorithm Example 1

T(1) = 1 for n=1 𝑛

T(n)¿ 𝑇 ( 2 )+ 1
𝑛
T(n) ¿ 𝑇 ( ) +1
( )
1 𝑛
2 ¿𝑇 +1
22
] +1
𝑛
¿𝑇 ( 𝑛
22 )+ 2 2 T(n)¿ 𝑇 (
2
)+ 1

] 𝑛
2 ¿𝑇 ( )
23 +
¿𝑇 ( )+3 𝑛
23
3

¿𝑇 (
2 𝑘 )+ k
𝑛
k
Time Complexity of Recursive Algorithm Example 1

T(1) = 1 for n=1

Assume
𝑛
T(n) ¿ 𝑇 ( ) +1 1
2 𝑛=2 𝑘
¿𝑇 ( )+ 2
𝑛
22 2
𝑛=2 𝑘
¿𝑇 (
2 3 )+3
𝑛
3 𝑙𝑜𝑔 𝑛=𝑙𝑜𝑔 2𝑘

¿𝑇 (
2 𝑘 )+ k
𝑛 𝑙𝑜𝑔 𝑛=k 𝑙𝑜𝑔 2 2
k
k
¿𝑇 ( +)log
𝑛
𝑛 n =1

+ log
¿ 1 n O (log n)
Time Complexity of Recursive Algorithm Example 2

T(1) = 1 for n=1 𝑛

T(n)¿ 𝑇 ( 2 )+ n
𝑛
T(n)
¿𝑇 ( )+ n

]
2 ¿𝑇 ( 𝑛
22 )
+
+n
+n 𝑛
T(n)¿ 𝑇 ( )+ n
2
( +n 𝑛
¿𝑇 ( )+
] ( +n 23

( ) +n
¿𝑇 ( 𝑛
23 )n ( ) + n
Time Complexity of Recursive Algorithm Example 2

T(1) = 1 for n=1

𝑛
T(n)¿ 𝑇 ( )+ n 1
2
n( )+n 2

𝟏 𝟏
¿𝑇 ( 𝑛
+ n
23 ) (
𝟐
)𝟐 + n (
𝟐
)+ n 3

-------------------------------

¿𝑇 ( 𝑛
2𝑘 +) n( )+] k
Time Complexity of Recursive Algorithm Example 2

T(1) = 1 for n=1 Assume

𝑛
T(n)¿ 𝑇 ( ) 𝑛=2 𝑘
2 +n
𝒂(𝟏 −𝒓 𝒌)
¿𝑇 ( 𝑛
+)
2𝑘 n] =
𝟏 −𝒓
𝟏𝒌 𝟏 =1
𝟏(𝟏 − ) 𝟏− 𝟏
¿𝑇 ( ) 𝑛
𝑛+ n. 𝟐𝒌
𝟏
𝟏−( )
¿𝑇 ( 1+) . n
𝟏
𝒏 r =
𝟐
𝟐 𝟐
𝒏−𝟏
¿ 1+ 2. n 𝒏 ¿ 1
+ 2 + 2 𝒏− 2
( 𝒏−𝟏 )¿ 1
¿2 𝒏 −𝟏 O (n)
Time Complexity of Recursive Algorithm Example 3

Consider the Algorithm Merge(A, p, q, r)

n1 = q - p + 1; n2 = r – q;
Let L[1 … n1+1] and R = [1 … n2+1] be new arrays
MERGE-SORT(A, p, r) { for i = 1 to n1
if p < r Dividing Cost
L [i] = A [p+i-1]
for j = 1 to n2
q = [(p+r)/2] R [j] = A [q+j]
MERGE-SORT(A, p, q) Two Recursive calls L [n1+1] = ∞
R [n2+1] = ∞
MERGE-SORT(A, q+1, r) each of size n/2 i=1 O(n)
j=1
Merge(A, p, q, r) for k = p to r
} Combine Step if L [i] ≤ R [j]
A [k] = L [i]
i = i+1
𝑛
T(n)¿ 𝟐 𝑇 ( )+ n else A [k] = R [j]
2 j = j+1

T(1) = 1 for n=1 f(n) denotes cost of the work done outside the recursive call, which includes:
 cost of dividing the problem
 cost of combining/merging the solutions
Time Complexity of Recursive Algorithm Example 3

T(1) = 1 for n=1 𝑛

T(n)¿ 2 𝑇 ( 2 )+ n
𝑛
T(n)¿ 𝟐 𝑇 ( )
2 +n
( )
1 𝑛
¿2𝑇 +
] 22
+n
+n 𝑛
T(n)¿ 2 𝑇 ( )+ n
2
+2n 2 𝑛
¿2𝑇 ( )+
] 23
2n

2n
¿𝟐𝟑𝑇 ( +) 3 n
𝑛
23 3
Time Complexity of Recursive Algorithm Example 3

T(1) = 1 for n=1

𝑛
T(n)¿ 𝟐 𝑇 ( )
2 +n 1

2n 2

¿𝟐𝟑𝑇 ( 𝑛
23 )3 n 3

-------------------------------

¿𝟐𝒌𝑇 ( 𝑛
2𝑘 )k n k
Time Complexity of Recursive Algorithm Example 3

T(1) = 1 for n=1

Assume
𝑛
T(n) ¿ 𝟐 𝑇 ( )
2 +n
𝑛=2 𝑘
¿𝟐𝒌𝑇( )k n 𝑛
2𝑘
𝑛=2 𝑘
¿𝒏𝑇(
𝑛 )n
𝑛 𝑙𝑜𝑔 2 𝑛=𝑙𝑜𝑔2 2 𝑘
𝑙𝑜𝑔 2 𝑛=k 𝑙𝑜𝑔 2 2
¿ 𝒏n k
O (n log n)
Time Complexity of Recursive Algorithm

References:

1.Appendix B: Solving Recurrence Equations, “Foundations of Algorithms” By Neapolitan (p-555)

2.Chapter 8: Advanced Counting Techniques, “Discrete Mathematics and Its Applications” By Rosen (p-501)

YouTube Link
https://www.youtube.com/watch?v=pauQm5VxnO8&list=LL&index=13