Subset Sum Queries in a Range using Bitset

• Difficulty Level : Medium
• Last Updated : 15 Dec, 2018

Given an array[] of N positive integers and M queries. Each query consists of two integers L and R represented by a range. For each query, find the count of numbers that lie in the given range which can be expressed as the sum of any subset of given array.

Prerequisite : Subset Sum Queries using Bitset
Examples:

Input : arr[] = { 1, 2, 2, 3, 5 }, M = 4
L = 1, R = 2
L = 1, R = 5
L = 3, R = 6
L = 9, R = 30
Output :
2
5
4
5
Explanation : For the first query, in range [1, 2] all numbers i.e. 1 and 2 can be expressed as a subset sum, 1 as 1, 2 as 2. For the second query, in range [1, 5] all numbers i.e. 1, 2, 3, 4 and 5 can be expressed as subset sum, 1 as 1, 2 as 2, 3 as 3, 4 as 2 + 2 or 1 + 3, 5 as 5. For the third query, in range [3, 6], all numbers i.e. 3, 4, 5 and 6 can be expressed as subset sum. For the last query, only numbers 9, 10, 11, 12, 13 can be expressed as subset sum, 9 as 5 + 2 + 2, 10 as 5 + 2 + 2 + 1, 11 as 5 + 3 + 2 + 1, 12 as 5 + 3 + 2 + 2 and 13 as 5 + 3 + 2 + 2 + 1.

Recommended: Please try your approach on {IDE} first, before moving on to the solution.

Approach : The idea is to use a bitset and iterate over the array to represent all possible subset sums. The current state of bitset is defined by ORing it with the previous state of bitset left shifted X times where X is the current element processed in the array. To answer the queries in O(1) time, we can precompute the count of numbers upto every number and for a range [L, R], answer would be pre[R] – pre[L – 1], where pre[] is the precomputed array.

Below is the implementation of above approach.

 // CPP Program to answer subset// sum queries in a given range#include using namespace std;  const int MAX = 1001;bitset bit;  // precomputation arrayint pre[MAX];  // structure to represent querystruct que {    int L, R;};  void answerQueries(int Q, que Queries[], int N,                   int arr[]){    // Setting bit at 0th position as 1    bit = 1;    for (int i = 0; i < N; i++)        bit |= (bit << arr[i]);      // Precompute the array    for (int i = 1; i < MAX; i++)        pre[i] = pre[i - 1] + bit[i];      // Answer Queries    for (int i = 0; i < Q; i++) {        int l = Queries[i].L;        int r = Queries[i].R;        cout << pre[r] - pre[l - 1] << endl;    }}  // Driver Code to test above functionint main(){    int arr[] = { 1, 2, 2, 3, 5 };    int N = sizeof(arr) / sizeof(arr);    int M = 4;    que Queries[M];    Queries.L = 1, Queries.R = 2;    Queries.L = 1, Queries.R = 5;    Queries.L = 3, Queries.R = 6;    Queries.L = 9, Queries.R = 30;    answerQueries(M, Queries, N, arr);    return 0;}
Output:
2
5
4
5

Time Complexity : Each query can be answered in O(1) time and precomputation requires O(MAX) time.

My Personal Notes arrow_drop_up