Number whose sum of XOR with given array range is maximum

Hello Everyone,

You are given a sequence of N integers and Q queries. In each query, you are given two parameters L and R. You have to find the smallest integer X such that 0 <= X < 2^31 and the sum of XOR of x with all elements is range [L, R] is maximum possible.

Examples :

Input : A = {20, 11, 18, 2, 13}
Three queries as (L, R) pairs
1 3
3 5
2 4
Output : 2147483629
2147483645
2147483645

Approach: The binary representation of each element and X, we can observe that each bit is independent and the problem can be solved by iterating over each bit. Now basically for each bit we need to count the number of 1’s and 0’s in the given range, if the number of 1’s are more then you have to set that bit of X to 0 so that the sum is maximum after xor with X else if number of 0’s are more then you have to set that bit of X to 1. If the number of 1’s and 0’s are equal then we can set that bit of X to any one of 1 or 0 because it will not affect the sum, but we have to minimize the value of X so we will take that bit 0.
Now, to optimize the solution we can pre-calculate the count of 1’s at each bit position of the numbers up to that position by making a prefix array this will take O(n) time. Now for each query number of 1’s will be the number of 1’s up to Rth position – number of 1’s up to (L-1)th position.

// CPP program to find smallest integer X

// such that sum of its XOR with range is

// maximum.

#include <bits/stdc++.h>

using namespace std;

#define MAX 2147483647

int one[100001][32];

// Function to make prefix array which

// counts 1's of each bit up to that number

void make_prefix( int A[], int n)

{

for ( int j = 0; j < 32; j++)

one[0][j] = 0;

// Making a prefix array which sums

// number of 1's up to that position

for ( int i = 1; i <= n; i++)

{

int a = A[i - 1];

for ( int j = 0; j < 32; j++)

{

int x = pow (2, j);

// If j-th bit of a number is set then

// add one to previously counted 1's

if (a & x)

one[i][j] = 1 + one[i - 1][j];

else

one[i][j] = one[i - 1][j];

}

}

}

// Function to find X

int Solve( int L, int R)

{

int l = L, r = R;

int tot_bits = r - l + 1;

// Initially taking maximum value all bits 1

int X = MAX;

// Iterating over each bit

for ( int i = 0; i < 31; i++)

{

// get 1's at ith bit between the

// range L-R by subtracting 1's till

// Rth number - 1's till L-1th number

int x = one[r][i] - one[l - 1][i];

// If 1's are more than or equal to 0's

// then unset the ith bit from answer

if (x >= tot_bits - x)

{

int ith_bit = pow (2, i);

// Set ith bit to 0 by doing

// Xor with 1

X = X ^ ith_bit;

}

}

return X;

}

// Driver program

int main()

{

// Taking inputs

int n = 5, q = 3;

int A[] = { 210, 11, 48, 22, 133 };

int L[] = { 1, 4, 2 }, R[] = { 3, 14, 4 };

make_prefix(A, n);

for ( int j = 0; j < q; j++)

cout << Solve(L[j], R[j]) << endl;

return 0;

}

Output :

2147483629
2147483647
2147483629