Algorithms

Lecture 5
Max Sum Subarray (Divide and Conquer)
Prantik Paul [PNP]
Lecturer
Department of Computer Science and
Engineering
1
BRAC University
 Maximum Subarray Problem
• Input: an array A[1..n] of n numbers
– Assume that some of the numbers are
negative, because this problem is trivial when
all numbers are nonnegative
• Output: a nonempty subarray A[i..j] having the
largest sum S[i, j] = ai + ai+1 +... + aj
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
A 13 -3 -25 20 -3 -16 -23 18 20 -7 12 -5 -22 15 -4 7

maximum subarray 2
 Maximum Subarray Problem
Target array 1 -4 3 2
: What is a
All the sub 1 1 maximum
arrays:
-4 -4 subarray?
Ans: The subarray
3 3
2
with the largest
2
-3
sum
1 -4
-4 3 -1 What is the
Max! 3 2 5 brute-force
1 -4 3 0 time?
-4 3 2 1
1 -4 3 2 2 3
 Brute-Force Algorithm
All possible contiguous subarrays
● A[1..1], A[1..2], A[1..3], ..., A[1..(n-1)], A[1..n]
● A[2..2], A[2..3], ..., A[2..(n-1)], A[2..n]
● ...
● A[(n-1)..(n-1)], A[(n-1)..n]
● A[n..n]

How many of them in total? O(n2)

Algorithm: For each subarray, compute the sum.

Find the subarray that has the maximum sum. 4
 Brute-Force Algorithm

Example: 2 -6 -1 3 -1
2 -2
sum from A[1]: 2 -4 -5 -2 -3 -1
-3
sum from A[2]: -6 -7 -4 -5 -3
-5
sum from A[3]: -1 2 1 3
1
sum from A[4]: 3 2 4
5
2
 Brute-Force Algorithm

Outer loop: index variable i to indicate start

of subarray, for 1 ≤ i ≤ n, i.e., A[1], A[2], ...,
A[n]
● for i = 1 to n do ...

Inner loop: for each start index i, we need to

go through A[i..i], A[i..(i+1)], ..., A[i..n]
● use an index j for i ≤ j ≤ n, i.e., consider
A[i..j]
● for j = i to n do ... 6
 Brute-Force Algorithm
max = -∞ Time
for i = 1 to n do complexity
begin ?
sum = 0 O(n2)
for j = i to n do
begin
sum = sum + A[j]
if sum > max
then max = sum
end
end
7
 Divide-and-Conquer Algorithm
Possible locations of a maximum subarray A[i..j] of
A[low..high], where mid = ⎣(low + high)/2⎦
▪ entirely in A[low..mid] (low ≤ i ≤ j ≤ mid)
▪ entirely in A[mid+1..high] (mid < i ≤ j ≤
high)
▪ crossing the midpoint (low ≤ i ≤ mid < j
crossing the midpoint
≤ high)
low mid
high

entirely in entirely in
A[low..mid] A[mid+1..high]
Possible locations of subarrays of A[low..high] 8
 Divide-and-Conquer Algorithm
FIND-MAX-CROSSING-SUBARRAY (A, low, mid, high)
left-sum = -∞ // Find a maximum subarray of the form A[i..mid]
sum = 0
for i = mid downto low
sum = sum + A[i ]
if sum > left-sum
left-sum = sum
max-left = i
right-sum = -∞ // Find a maximum subarray of the form A[mid + 1 .. j ]
sum =0
for j = mid +1 to high
sum = sum + A[j]
if sum > right-sum
right-sum = sum
max-right = j
// Return the indices and the sum of the two subarrays
9
return (max-left, max-right, left-sum + right-sum)
 Divide-and-Conquer Algorithm

A[mid+1..j]
low i mid
high
mid +1 j
A[i..mi
d]
A[i..j] comprises two subarrays A[i..mid] and
A[mid+1..j]

10
 Divide-and-Conquer Algorithm
mid =5
1 2 3 4 5 6 7 8 9 10
13 -3 -25 20 -3 -16 -23 18 20 -7
A
S[5 .. 5] =
-3
S[4 .. 5] =
17 ⇐ (max-left = 4)
S[3 .. 5] = mid-8=5
S[2 .. 5] = 1 2 3 -11
4 5 6 7 8 9 10
S[1 .. 5] = 13 -3 2 -25 20 -3 -16 -23 18 20 -7
A
S[6 .. 6] = -16
S[6 .. 7] = -39
S[6 .. 8] = -21
S[6 .. 9] = (max-right = 9) ⇒ -
1
S[6..10] = - 11
⇒ maximum subarray crossing
8 mid is S[4..9] = 16
 Divide-and-Conquer Algorithm
FIND-MAXIMUM-SUBARRAY (A, low, high)
if high == low
return (low, high, A[low]) // base case: only one element
else mid =
(left-low, left-high, left-sum) =
FIND-MAXIMUM-SUBARRAY(A, low, mid)
(right-low, right-high, right-sum) =
FIND-MAXIMUM-SUBARRAY(A, mid + 1, high)
(cross-low, cross-high, cross-sum) =
FIND-MAX-CROSSING-SUBARRAY(A, low, mid, high)
if left-sum ≧ right-sum and left-sum ≧ cross-sum
return (left-low, left-high, left-sum)
elseif right-sum ≧ left-sum and right-sum ≧ cross-sum
return (right-low, right-high, right-sum)
else return (cross-low, cross-high, cross-sum)
Initial call: FIND-MAXIMUM-SUBARRAY (A, 1, n) 12
 Divide-and-Conquer Algorithm
Analyzing time complexity
FIND-MAX-CROSSING-SUBARRAY : Θ(n), where n = high
− low + 1

FIND-MAXIMUM-SUBARRAY

T(n) = 2T(n/2) + Θ(n)

= Θ(n lg n) (similar to merge-sort) 13
Practice

-2 1 -3 4 -1 2 1 -5 4
Resources

● CLRS Book:
○ Chapter 4.1 - The Maximum Subarray Problem

15