Maximize sum of K elements selected from a Matrix such that each selected element must be preceded by selected row elements

Given a 2D array arr[][] of size N * M, and an integer K, the task is to select K elements with maximum possible sum such that if an element arr[i][j] is selected, then all the elements from the i^th row present before the j^th column needs to be selected.

Examples:

Input: arr[][] = {{10, 10, 100, 30}, {80, 50, 10, 50}}, K = 5
Output: 250
Explanation:
Selecting first 3 elements from the first row, sum = 10 + 10 + 100 = 120
Selecting first 2 elements from the second row, sum = 80 + 50 = 130
Therefore, the maximum sum = 120 + 130 = 250.

Input: arr[][] = {{30, 10, 110, 13}, {810, 152, 12, 5}, {124, 24, 54, 124}}, K = 6
Output: 1288
Explanation:
Selecting first 2 elements from the second row, sum = 810 + 152 = 962
Selecting all 4 elements from the third row, sum = 124 + 24 + 54 + 124 = 326
Therefore, the maximum sum = 962 + 326 = 1288

Naive Approach: The simplest approach to solve the problem is to generate all possible subsets of size K from the given matrix based on specified constraint and calculate the maximum sum possible from these subsets. 

Time Complexity: O((N*M)!/(K!)*(N*M – K)!)
Auxiliary Space: O(1)

Efficient Approach: The above approach can be optimized by using Dynamic Programming. The idea is to find the maximum sum of selecting i elements till the j^th row of the 2D array arr[][]. The auxiliary array dp[i][j + 1] stores the maximum sum of selecting i elements till j^th row of matrix arr[][]. Follow the steps below to solve the problem:

• Initialize a dp[][] table of size (K+1)*(N+1) and initialize dp[0][i] as 0, as no elements have sum equal to 0.
• Initialize dp[i][0] as 0, as there are no elements to be selected.
• Iterate over the range [0, K] using two nested loops and [0, N] using the variables i and j respectively:
  • Initialize a variable, say sum as 0, to keep track of the cumulative sum of elements in the j^th row of arr[][].
  • Initialize maxSum as dp[i][j], if no item needs to be selected from j^th row of arr[][].
  • Iterate with k as 1 until k > M or k > i:
    • Increment sum += buy[j][k – 1].
    • Store the maximum of existing maxSum and (sum + dp[i – k][j]) in maxSum.
  • Store dp[i][j + 1] as the value of maxSum.
• Print the value dp[K][N] as the maximum sum of K elements till (N – 1)^th row of matrix arr[][].

Below is the implementation of the above approach:

C++

// C++ program for the above approach
#include <iostream>
using namespace std;

// Function to return the maximum
// of two elements
int max(int a, int b)
{
    return a > b ? a : b;
}

// Function to find the maximum sum
// of selecting K elements from
// the given 2D array arr[][]
int maximumsum(int arr[][4], int K,
               int N, int M)
{
    int sum = 0, maxSum;
    int i, j, k;

    // dp table of size (K+1)*(N+1)
    int dp[K + 1][N + 1];

    // Initialize dp[0][i] = 0
    for (i = 0; i <= N; i++)
        dp[0][i] = 0;

    // Initialize dp[i][0] = 0
    for (i = 0; i <= K; i++)
        dp[i][0] = 0;

    // Selecting i elements
    for (i = 1; i <= K; i++) {

        // Select i elements till jth row
        for (j = 0; j < N; j++) {

            // dp[i][j+1] is the maximum
            // of selecting i elements till
            // jth row

            // sum = 0, to keep track of
            // cumulative elements sum
            sum = 0;
            maxSum = dp[i][j];

            // Traverse arr[j][k] until
            // number of elements until k>i
            for (k = 1; k <= M && k <= i; k++) {

                // Select arr[j][k - 1]th item
                sum += arr[j][k - 1];

                maxSum
                    = max(maxSum,
                          sum + dp[i - k][j]);
            }

            // Store the maxSum in dp[i][j+1]
            dp[i][j + 1] = maxSum;
        }
    }

    // Return the maximum sum
    return dp[K][N];
}

// Driver Code
int main()
{
    int arr[][4] = { { 10, 10, 100, 30 },
                     { 80, 50, 10, 50 } };

    int N = 2, M = 4;
    int K = 5;

    // Function Call
    cout << maximumsum(arr, K, N, M);

    return 0;
}

Java

// Java program for the above approach
import java.util.*;

class GFG{

// Function to return the maximum
// of two elements
static int max(int a, int b)
{
    return a > b ? a : b;
}

// Function to find the maximum sum
// of selecting K elements from
// the given 2D array arr[][]
static int maximumsum(int arr[][], int K,
                      int N, int M)
{
    int sum = 0, maxSum;
    int i, j, k;

    // dp table of size (K+1)*(N+1)
    int[][] dp = new int[K + 1][N + 1];

    // Initialize dp[0][i] = 0
    for(i = 0; i <= N; i++)
        dp[0][i] = 0;

    // Initialize dp[i][0] = 0
    for(i = 0; i <= K; i++)
        dp[i][0] = 0;

    // Selecting i elements
    for(i = 1; i <= K; i++)
    {

        // Select i elements till jth row
        for(j = 0; j < N; j++)
        {

            // dp[i][j+1] is the maximum
            // of selecting i elements till
            // jth row

            // sum = 0, to keep track of
            // cumulative elements sum
            sum = 0;
            maxSum = dp[i][j];

            // Traverse arr[j][k] until
            // number of elements until k>i
            for(k = 1; k <= M && k <= i; k++)
            {

                // Select arr[j][k - 1]th item
                sum += arr[j][k - 1];

                maxSum = Math.max(maxSum,
                                  sum + dp[i - k][j]);
            }

            // Store the maxSum in dp[i][j+1]
            dp[i][j + 1] = maxSum;
        }
    }

    // Return the maximum sum
    return dp[K][N];
}

// Driver Code
public static void main(String[] args)
{
    int arr[][] =  { { 10, 10, 100, 30 },
                     { 80, 50, 10, 50 } };

    int N = 2, M = 4;
    int K = 5;

    // Function Call
    System.out.print(maximumsum(arr, K, N, M));
}
}

// This code is contributed by susmitakundugoaldanga

Python3

# Python program for the above approach
import math;

# Function to return the maximum
# of two elements
def max(a, b):
    if(a > b):
        return a;
    else:
        return b;

# Function to find the maximum sum
# of selecting K elements from
# the given 2D array arr
def maximumsum(arr, K, N, M):
    sum = 0;
    maxSum = 0;

    # dp table of size (K+1)*(N+1)
    dp = [[0 for i in range(N + 1)] for j in range(K + 1)]

    # Initialize dp[0][i] = 0
    for i in range(0, N + 1):
        dp[0][i] = 0;

    # Initialize dp[i][0] = 0
    for i in range(0, K + 1):
        dp[i][0] = 0;

    # Selecting i elements
    for  i in range(1, K + 1):

        # Select i elements till jth row
        for  j in range(0, N):

            # dp[i][j+1] is the maximum
            # of selecting i elements till
            # jth row

            # sum = 0, to keep track of
            # cumulative elements sum
            sum = 0;
            maxSum = dp[i][j];

            # Traverse arr[j][k] until
            # number of elements until k>i
            for k in range(1, i + 1):
                if(k > M):
                    break;

                # Select arr[j][k - 1]th item
                sum += arr[j][k - 1];

                maxSum = max(maxSum, sum + dp[i - k][j]);

            # Store the maxSum in dp[i][j+1]
            dp[i][j + 1] = maxSum;

    # Return the maximum sum
    return dp[K][N];

# Driver Code
if __name__ == '__main__':
    arr = [[10, 10, 100, 30], [80, 50, 10, 50]];

    N = 2;
    M = 4;
    K = 5;

    # Function Call
    print(maximumsum(arr, K, N, M));


    # This code is contributed by 29AjayKumar

C#

// C# program for the above approach
using System;

class GFG{

// Function to return the maximum
// of two elements
static int max(int a, int b)
{
    return a > b ? a : b;
}

// Function to find the maximum sum
// of selecting K elements from
// the given 2D array [,]arr
static int maximumsum(int [,]arr, int K,
                      int N, int M)
{
    int sum = 0, maxSum;
    int i, j, k;

    // dp table of size (K+1)*(N+1)
    int[,] dp = new int[K + 1, N + 1];

    // Initialize dp[0,i] = 0
    for(i = 0; i <= N; i++)
        dp[0, i] = 0;

    // Initialize dp[i,0] = 0
    for(i = 0; i <= K; i++)
        dp[i, 0] = 0;

    // Selecting i elements
    for(i = 1; i <= K; i++)
    {

        // Select i elements till jth row
        for(j = 0; j < N; j++)
        {

            // dp[i,j+1] is the maximum
            // of selecting i elements till
            // jth row

            // sum = 0, to keep track of
            // cumulative elements sum
            sum = 0;
            maxSum = dp[i, j];

            // Traverse arr[j,k] until
            // number of elements until k>i
            for(k = 1; k <= M && k <= i; k++)
            {

                // Select arr[j,k - 1]th item
                sum += arr[j, k - 1];

                maxSum = Math.Max(maxSum,
                                  sum + dp[i - k, j]);
            }

            // Store the maxSum in dp[i,j+1]
            dp[i, j + 1] = maxSum;
        }
    }

    // Return the maximum sum
    return dp[K, N];
}

// Driver Code
public static void Main(String[] args)
{
    int [,]arr =  { { 10, 10, 100, 30 },
                    { 80, 50, 10, 50 } };

    int N = 2, M = 4;
    int K = 5;

    // Function Call
    Console.Write(maximumsum(arr, K, N, M));
}
}

// This code is contributed by Amit Katiyar

Javascript

<script>
// Javascript program to implement
// the above approach

// Function to return the maximum
// of two elements
function max(a, b)
{
    return a > b ? a : b;
}

// Function to find the maximum sum
// of selecting K elements from
// the given 2D array arr[][]
function maximumsum(arr, K,
                      N, M)
{
    let sum = 0, maxSum;
    let i, j, k;

    // dp table of size (K+1)*(N+1)
    let dp = new Array(K + 1);

    // Loop to create 2D array using 1D array
    for ( i = 0; i < dp.length; i++) {
        dp[i] = new Array(2);
    }

    // Initialize dp[0][i] = 0
    for(i = 0; i <= N; i++)
        dp[0][i] = 0;

    // Initialize dp[i][0] = 0
    for(i = 0; i <= K; i++)
        dp[i][0] = 0;

    // Selecting i elements
    for(i = 1; i <= K; i++)
    {

        // Select i elements till jth row
        for(j = 0; j < N; j++)
        {

            // dp[i][j+1] is the maximum
            // of selecting i elements till
            // jth row

            // sum = 0, to keep track of
            // cumulative elements sum
            sum = 0;
            maxSum = dp[i][j];

            // Traverse arr[j][k] until
            // number of elements until k>i
            for(k = 1; k <= M && k <= i; k++)
            {

                // Select arr[j][k - 1]th item
                sum += arr[j][k - 1];

                maxSum = Math.max(maxSum,
                                  sum + dp[i - k][j]);
            }

            // Store the maxSum in dp[i][j+1]
            dp[i][j + 1] = maxSum;
        }
    }

    // Return the maximum sum
    return dp[K][N];
}

// Driver code
    let arr =  [[ 10, 10, 100, 30 ],
                     [ 80, 50, 10, 50 ]];

    let N = 2, M = 4;
    let K = 5;

    // Function Call
    document.write(maximumsum(arr, K, N, M));

 // This code is contributed by souravghosh0416.
</script>

Output:

250

Time Complexity: O(K*N*M)
Auxiliary Space: O(K*N)