QUESTION :
Given weights and values of N items, we need to put these items in a knapsack of capacity W to get the maximum total value in the knapsack.
Note: Unlike 0/1 knapsack, you are allowed to break the item.

THOUGHT PROCESS
The logic is simple , sort all the items on basis of 'VALUE PER UNIT' and add whole item if it's weight is less than capacity otherwise
add the portion of that item.

ALGORITHM

1. Declare a max priority queue and implement its comparitor based on VALUE PER UNIT
   2) Now push every item in pq
   3. While (capacity > 0 && pq is not empty) keep poping the top element
      3.1) If the top item's weight <= capacity
      3.1.1) add it's value in 'total_value' and decrease the capacity by item's weight
      3.2) else if item's weight > capacity
      3.2.1) increment value accoring to the remaning capacity only and make capacity to zero
      ==> total_value += (double)top.value / top.weight * capacity;
   4. return the total_value

CODE

struct Item
{
    int value;
    int weight;
};
struct comparitor_
{
    bool operator()(Item min, Item max)
    {

        return (double)max.value / max.weight > (double)min.value / min.weight;
    }
};
double fractionalKnapsack(int capacity, Item arr[], int n)
{
    priority_queue<Item, vector<Item>, comparitor_> pq;
    for (int i = 0; i < n; i++)
    {
        pq.push(arr[i]);
    }

    double total_value = 0;
    while (capacity && pq.size())
    {
        auto top = pq.top();
        pq.pop();
        if (top.weight <= capacity)
        {
            total_value += top.value;
            capacity -= top.weight;
        }
        else
        {
            total_value += (double)top.value / top.weight * capacity;
            capacity = 0;
        }
    }
    return total_value;
}