I am trying to solve this question:

Let's say you have a binary heap and an index $i$, design an algorithm that finds if a number $x$ appears in the path between the root of the heap and $heap[i]$ in time $O(\log(\log n))$.

My solution:

Before the loop, I will create a new linked list and initiate $heap[i]$ to be the first item in the list. Then, I will run a loop on the heap that will start from index $i$, and in every iteration, it will extract the parent of the item in the index $i$ to a new array and will break after extracting the root of the heap.

Then, I will use binary search on the new array (I know that every parent in the heap is bigger than its children. So the new array is already sorted). The problem is that the extraction of the parents cost me $O(\log i)$ which can be in the worst-case $O(\log n)$, and for that reason, my solution can't work.

I will be grateful for any help.


1 Answer 1


Note that a path from the root to $heap[i]$ is in the sorted order. Moreover, the path length is $O(\log n)$. You just need to do binary search on this path. For that, you need to access the $i^{th}$ ancestor of any node in the heap in $O(1)$ time. Assuming $1$ based indexing, the $1^{st}$ ancestor of $heap[i]$ is located at $\lfloor \frac{i}{2} \rfloor$ location in the array. The $2^{nd}$ ancestor of $heap[i]$ is located at $\lfloor \frac{i}{4} \rfloor$ location, and so on. Using it you can easily perform binary search on the path from root to $heap[i]$ in $O(\log \log n)$ time.


Assume that we have a min heap. Initialize, start = $1$ and end = $h = \lfloor \log(i) \rfloor$

BinarySerach (start, end)

mid = (start + end)/2
index of mid = i/(2^(h- mid)) //(h-mid)^th ancestor of i

if (end-start = 1)
    if (heap[index of mid] == x) return "YES"; otherwise "NO".

if (heap[index of mid] >= x)
   return BinarySearch (start, mid)
   return BinarySearch (mid+1,end)
  • $\begingroup$ Hi, thanks for your answer. Even that i understand your answer, I have been struggling to develop this alghoritem you suggest , if you can write a pseudo code that doing this binary se search it will help me alot. $\endgroup$
    – yuval
    May 9, 2021 at 9:51
  • 1
    $\begingroup$ @yuval Can you tell where are you struggling? I can help with that. $\endgroup$ May 9, 2021 at 14:53
  • $\begingroup$ I tried to wrote a pseudo code and i can't see how to devide the path items each time (like in binary search where you dividing by 2). I mean, its a little long to explain everything i tried and i will be glad if you will just write for me a pseudo code of the iteration it will be less time for me and less time for you. $\endgroup$
    – yuval
    May 9, 2021 at 14:57
  • $\begingroup$ I know it's important to solve things yourself but i solved a harder problems for sure i am just missing something and maybe see the solution is the right thing to do after putting somework on this thing $\endgroup$
    – yuval
    May 9, 2021 at 14:58
  • 1
    $\begingroup$ @yuval I have added the pseudocode. Let me know if there is any confusion. $\endgroup$ May 9, 2021 at 16:46

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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