We are given a binary max (every node is larger than its children) heap with $n$ elements.

We now change $\frac{n}{4}$ of the elements at random. We don't know which ones and to which value. And so, it is now a possibility that the heap is no longer a valid max heap.

Show that the minimum number of steps required to fix this into a max heap is $\Theta (n)$.

I realize why this can't be lower than $n$. after all, we need to access every element in the heap to check if it was changed or not, it can't be less then $n$, but why is the fastest way $n$? I can't think of such an algorithm.

The algorithm i came up with, which is perhaps the naive one, is to insert all of the values of the heap to an array, and sorting it. but that is $\Theta (n \log n)$


Managed to solve it.

There is an algorithm to build a heap in $\Theta (n)$ time called buildheap.

  • 1
    $\begingroup$ Could you add a link to that algorithm here, so that other users having the same problem can have an answer? Thank you. $\endgroup$ – Hippalectryon Oct 4 '14 at 11:24
  • $\begingroup$ The algorithm is recursive: Starting from the root, buildheap(leftchild), then buildheap(rightchild), then heapify the root. By master's theorem that runs in O(n) $\endgroup$ – Oria Gruber Oct 18 '14 at 13:21

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.