Open In App

Maximum Subarray Sum using Divide and Conquer algorithm

Improve
Improve
Improve
Like Article
Like
Save Article
Save
Share
Report issue
Report
 

You are given a one dimensional array that may contain both positive and negative integers, find the sum of contiguous subarray of numbers which has the largest sum.

For example, if the given array is {-2, -5, 6, -2, -3, 1, 5, -6}, then the maximum subarray sum is 7 (see highlighted elements).

The naive method is to run two loops. The outer loop picks the beginning element, the inner loop finds the maximum possible sum with first element picked by outer loop and compares this maximum with the overall maximum. Finally, return the overall maximum. The time complexity of the Naive method is O(n^2).

Using Divide and Conquer approach, we can find the maximum subarray sum in O(nLogn) time. Following is the Divide and Conquer algorithm. 

  1. Divide the given array in two halves
  2. Return the maximum of following three
    • Maximum subarray sum in left half (Make a recursive call)
    • Maximum subarray sum in right half (Make a recursive call)
    • Maximum subarray sum such that the subarray crosses the midpoint

The lines 2.a and 2.b are simple recursive calls. How to find maximum subarray sum such that the subarray crosses the midpoint? We can easily find the crossing sum in linear time. The idea is simple, find the maximum sum starting from mid point and ending at some point on left of mid, then find the maximum sum starting from mid + 1 and ending with some point on right of mid + 1. Finally, combine the two and return the maximum among left, right and combination of both.

Below is the implementation of the above approach: 

C++




#include <bits/stdc++.h>
using namespace std;
  
// A utility function to find maximum of two integers
int max(int a, int b) {
    return (a > b) ? a : b;
}
  
// A utility function to find maximum of three integers
int max(int a, int b, int c) {
    return max(max(a, b), c);
}
  
// Find the maximum possible sum in arr[] such that arr[m]
// is part of it
int maxCrossingSum(int arr[], int l, int m, int h) {
    // Include elements on left of mid.
    int sum = 0;
    int left_sum = INT_MIN;
    for (int i = m; i >= l; i--) {
        sum = sum + arr[i];
        if (sum > left_sum)
            left_sum = sum;
    }
  
    // Include elements on right of mid
    sum = 0;
    int right_sum = INT_MIN;
    for (int i = m; i <= h; i++) {
        sum = sum + arr[i];
        if (sum > right_sum)
            right_sum = sum;
    }
  
    // Return sum of elements on left and right of mid
    // returning only left_sum + right_sum will fail for
    // [-2, 1]
    return max(left_sum + right_sum - arr[m], left_sum, right_sum);
}
  
// Returns sum of maximum sum subarray in aa[l..h]
int maxSubArraySum(int arr[], int l, int h) {
    // Invalid Range: low is greater than high
    if (l > h)
        return INT_MIN;
    // Base Case: Only one element
    if (l == h)
        return arr[l];
  
    // Find middle point
    int m = (l + h) / 2;
  
    /* Return maximum of following three possible cases
            a) Maximum subarray sum in left half
            b) Maximum subarray sum in right half
            c) Maximum subarray sum such that the subarray
       crosses the midpoint */
    return max(maxSubArraySum(arr, l, m - 1),
               maxSubArraySum(arr, m + 1, h),
               maxCrossingSum(arr, l, m, h));
}
  
/*Driver program to test maxSubArraySum*/
int main() {
    int arr[] = { 2, 3, 4, 5, 7 };
    int n = sizeof(arr) / sizeof(arr[0]);
    int max_sum = maxSubArraySum(arr, 0, n - 1);
    cout << "Maximum contiguous sum is " << max_sum;
    return 0;
}


Java




// A Divide and Conquer based Java
// program for maximum subarray sum
// problem
import java.util.*;
  
class GFG {
  
    // Find the maximum possible sum in arr[]
    // such that arr[m] is part of it
    static int maxCrossingSum(int arr[], int l, int m,
                              int h)
    {
        // Include elements on left of mid.
        int sum = 0;
        int left_sum = Integer.MIN_VALUE;
        for (int i = m; i >= l; i--) {
            sum = sum + arr[i];
            if (sum > left_sum)
                left_sum = sum;
        }
  
        // Include elements on right of mid
        sum = 0;
        int right_sum = Integer.MIN_VALUE;
        for (int i = m; i <= h; i++) {
            sum = sum + arr[i];
            if (sum > right_sum)
                right_sum = sum;
        }
  
        // Return sum of elements on left
        // and right of mid
        // returning only left_sum + right_sum will fail for
        // [-2, 1]
        return Math.max(left_sum + right_sum - arr[m],
                        Math.max(left_sum, right_sum));
    }
  
    // Returns sum of maximum sum subarray
    // in aa[l..h]
    static int maxSubArraySum(int arr[], int l, int h)
    {
          //Invalid Range: low is greater than high
          if (l > h)
              return Integer.MIN_VALUE;
        // Base Case: Only one element
        if (l == h)
            return arr[l];
  
        // Find middle point
        int m = (l + h) / 2;
  
        /* Return maximum of following three
        possible cases:
        a) Maximum subarray sum in left half
        b) Maximum subarray sum in right half
        c) Maximum subarray sum such that the
        subarray crosses the midpoint */
        return Math.max(
            Math.max(maxSubArraySum(arr, l, m-1),
                     maxSubArraySum(arr, m + 1, h)),
            maxCrossingSum(arr, l, m, h));
    }
  
    /* Driver program to test maxSubArraySum */
    public static void main(String[] args)
    {
        int arr[] = { 2, 3, 4, 5, 7 };
        int n = arr.length;
        int max_sum = maxSubArraySum(arr, 0, n - 1);
  
        System.out.println("Maximum contiguous sum is "
                           + max_sum);
    }
}
// This code is contributed by Prerna Saini


Python




# A Divide and Conquer based program
# for maximum subarray sum problem
  
# Find the maximum possible sum in
# arr[] auch that arr[m] is part of it
  
  
def maxCrossingSum(arr, l, m, h):
  
    # Include elements on left of mid.
    sm = 0
    left_sum = -10000
  
    for i in range(m, l-1, -1):
        sm = sm + arr[i]
  
        if (sm > left_sum):
            left_sum = sm
  
    # Include elements on right of mid
    sm = 0
    right_sum = -1000
    for i in range(m, h + 1):
        sm = sm + arr[i]
  
        if (sm > right_sum):
            right_sum = sm
  
    # Return sum of elements on left and right of mid
    # returning only left_sum + right_sum will fail for [-2, 1]
    return max(left_sum + right_sum - arr[m], left_sum, right_sum)
  
  
# Returns sum of maximum sum subarray in aa[l..h]
def maxSubArraySum(arr, l, h):
    #Invalid Range: low is greater than high
    if (l > h):
        return -10000
    # Base Case: Only one element
    if (l == h):
        return arr[l]
  
    # Find middle point
    m = (l + h) // 2
  
    # Return maximum of following three possible cases
    # a) Maximum subarray sum in left half
    # b) Maximum subarray sum in right half
    # c) Maximum subarray sum such that the
    #     subarray crosses the midpoint
    return max(maxSubArraySum(arr, l, m-1),
               maxSubArraySum(arr, m+1, h),
               maxCrossingSum(arr, l, m, h))
  
  
# Driver Code
arr = [2, 3, 4, 5, 7]
n = len(arr)
  
max_sum = maxSubArraySum(arr, 0, n-1)
print("Maximum contiguous sum is ", max_sum)
  
# This code is contributed by Nikita Tiwari.


C#




// A Divide and Conquer based C#
// program for maximum subarray sum
// problem
using System;
  
class GFG {
  
    // Find the maximum possible sum in arr[]
    // such that arr[m] is part of it
    static int maxCrossingSum(int[] arr, int l, int m,
                              int h)
    {
        // Include elements on left of mid.
        int sum = 0;
        int left_sum = int.MinValue;
        for (int i = m; i >= l; i--) {
            sum = sum + arr[i];
            if (sum > left_sum)
                left_sum = sum;
        }
  
        // Include elements on right of mid
        sum = 0;
        int right_sum = int.MinValue;
        ;
        for (int i = m; i <= h; i++) {
            sum = sum + arr[i];
            if (sum > right_sum)
                right_sum = sum;
        }
  
        // Return sum of elements on left
        // and right of mid
        // returning only left_sum + right_sum will fail for
        // [-2, 1]
        return Math.Max(left_sum + right_sum -arr[m],
                        Math.Max(left_sum, right_sum));
    }
  
    // Returns sum of maximum sum subarray
    // in aa[l..h]
    static int maxSubArraySum(int[] arr, int l, int h)
    {
        //Invalid Range: low is greater than high
          if (l > h)
              return int.MinValue;
        // Base Case: Only one element
        if (l == h)
            return arr[l];
  
        // Find middle point
        int m = (l + h) / 2;
  
        /* Return maximum of following three
        possible cases:
        a) Maximum subarray sum in left half
        b) Maximum subarray sum in right half
        c) Maximum subarray sum such that the
        subarray crosses the midpoint */
        return Math.Max(
            Math.Max(maxSubArraySum(arr, l, m-1),
                     maxSubArraySum(arr, m + 1, h)),
            maxCrossingSum(arr, l, m, h));
    }
  
    /* Driver program to test maxSubArraySum */
    public static void Main()
    {
        int[] arr = { 2, 3, 4, 5, 7 };
        int n = arr.Length;
        int max_sum = maxSubArraySum(arr, 0, n - 1);
  
        Console.Write("Maximum contiguous sum is "
                      + max_sum);
    }
}
  
// This code is contributed by vt_m.


PHP




<?php 
// A Divide and Conquer based program 
// for maximum subarray sum problem 
  
// Find the maximum possible sum in arr[] 
// such that arr[m] is part of it 
function maxCrossingSum(&$arr, $l, $m, $h
    // Include elements on left of mid. 
    $sum = 0; 
    $left_sum = PHP_INT_MIN; 
    for ($i = $m; $i >= $l; $i--) 
    
        $sum = $sum + $arr[$i]; 
        if ($sum > $left_sum
        $left_sum = $sum
    
  
    // Include elements on right of mid 
    $sum = 0; 
    $right_sum = PHP_INT_MIN; 
    for ($i = $m; $i <= $h; $i++) 
    
        $sum = $sum + $arr[$i]; 
        if ($sum > $right_sum
        $right_sum = $sum
    
  
    // Return sum of elements on left 
    // and right of mid 
    // returning only left_sum + right_sum will fail for [-2, 1] 
    return max($left_sum + $right_sum - $arr[$m], $left_sum, $right_sum); 
  
// Returns sum of maximum sum 
// subarray in aa[l..h] 
function maxSubArraySum(&$arr, $l, $h
      //Invalid Range: low is greater than high
      if ($l > $h)
          return PHP_INT_MIN;
    
    // Base Case: Only one element 
    if ($l == $h
        return $arr[$l]; 
      
    // Find middle point 
    $m = intval(($l + $h) / 2); 
      
    /* Return maximum of following three possible cases 
        a) Maximum subarray sum in left half 
        b) Maximum subarray sum in right half 
        c) Maximum subarray sum such that the 
        subarray crosses the midpoint */
    return max(maxSubArraySum($arr, $l, $m-1), 
            maxSubArraySum($arr, $m + 1, $h), 
            maxCrossingSum($arr, $l, $m, $h)); 
  
// Driver Code 
$arr = array(2, 3, 4, 5, 7); 
$n = count($arr); 
$max_sum = maxSubArraySum($arr, 0, $n - 1); 
echo "Maximum contiguous sum is " . $max_sum
  
// This code is contributed by rathbhupendra, Aadil 
?> 


Javascript




<script>
    // A Divide and Conquer based program for maximum subarray
    // sum problem
  
    // A utility function to find maximum of two integers
    function max(a,b) { return (a > b) ? a : b; }
  
    // A utility function to find maximum of three integers
    function max(a,b,c) { return Math.max(Math.max(a, b), c); }
  
    // Find the maximum possible sum in arr[] auch that arr[m]
    // is part of it
    function maxCrossingSum(arr, l, m,h)
    {
        // Include elements on left of mid.
        let sum = 0;
        let left_sum = Number.MIN_VALUE;
        for (let i = m; i >= l; i--) {
            sum = sum + arr[i];
            if (sum > left_sum)
                left_sum = sum;
        }
  
        // Include elements on right of mid
        sum = 0;
        let right_sum = Number.MIN_VALUE;
        for (let i = m; i <= h; i++) {
            sum = sum + arr[i];
            if (sum > right_sum)
                right_sum = sum;
        }
  
        // Return sum of elements on left and right of mid
        // returning only left_sum + right_sum will fail for
        // [-2, 1]
        return max(left_sum + right_sum - arr[m], left_sum, right_sum);
    }
  
    // Returns sum of maximum sum subarray in aa[l..h]
    function maxSubArraySum(arr, l,h)
    {
        //Invalid Range: low is greater than high
          if (l > h)
              return Number.MIN_VALUE;
          
        // Base Case: Only one element
        if (l == h)
            return arr[l];
  
        // Find middle point
        let m = parseInt((l + h) / 2, 10);
  
        /* Return maximum of following three possible cases
                a) Maximum subarray sum in left half
                b) Maximum subarray sum in right half
                c) Maximum subarray sum such that the subarray
        crosses the midpoint */
        return max(maxSubArraySum(arr, l, m-1),
                maxSubArraySum(arr, m + 1, h),
                maxCrossingSum(arr, l, m, h));
    }
  
  
    let arr = [ 2, 3, 4, 5, 7 ];
    let n = arr.length;
    let max_sum = maxSubArraySum(arr, 0, n - 1);
    document.write("Maximum contiguous sum is " + max_sum);
      
  // This code is contributed by vaibhavrabadiya117.  
</script>


Output

Maximum contiguous sum is 21

Time Complexity: maxSubArraySum() is a recursive method and time complexity can be expressed as following recurrence relation. 
T(n) = 2T(n/2) + ?(n) 

Time Complexity : O(nlogn)

Auxiliary Space: O(1).
The above recurrence is similar to Merge Sort and can be solved either using Recurrence Tree method or Master method. It falls in case II of Master Method and solution of the recurrence is ?(nLogn). 

The Kadane’s Algorithm for this problem takes O(n) time. Therefore the Kadane’s algorithm is better than the Divide and Conquer approach, but this problem can be considered as a good example to show power of Divide and Conquer. The above simple approach where we divide the array in two halves, reduces the time complexity from O(n^2) to O(nLogn).

 



Last Updated : 18 Sep, 2023
Like Article
Save Article
Previous
Next
Share your thoughts in the comments
Similar Reads