Is O(n * sqrt(n * a (n)) possible?

Revision en1, by rand, 2024-12-04 20:39:27

Hi codeforces!

I recently wrote sqrt-decomposition + dsu, and I only had to use snm if the query contains the whole block. What would be the asymptotics of this solution?

It may seem obvious to you that the asymptotics will be $$$O(n \cdot \sqrt{n \cdot \varphi^{-1}} {(n)}})$$$ ($$$\varphi^{-1} {(n)}$$$ — the inverse Ackerman function).

But in fact the asymptotics is — $$$O(n \cdot B \cdot \varphi^{-1}{(n)} + \frac{n ^ 2}{B})$$$ ($$$B$$$ — block size) and in order to achieve the asymptotics with an accreman under the sqrt we need to know $$$\varphi^{-1}{(n)}$$$ to put $$$B = \sqrt{n * \varphi^{-1}{-1} {(n)}}$$$, and to do that we need to know $$$\varphi^{-1} {(n)}$$$. So it comes down to a quick calculation of $$$\varphi^{-1} {(n)}$$$. Is it possible to do this?

Tags dsu, sqrt_decomposition, asymptotics, theoretical

History

 
 
 
 
Revisions
 
 
  Rev. Lang. By When Δ Comment
en2 English rand 2024-12-04 20:40:28 1 Tiny change: 'arphi^{-1}} {(n)}})$' -> 'arphi^{-1} {(n)}})$'
en1 English rand 2024-12-04 20:39:27 803 Initial revision for English translation
ru2 Russian rand 2024-12-04 20:35:18 15 Мелкая правка: 'жно знать a(n), чтобы по' -> 'жно знать $\varphi^{-1}{(n)}$, чтобы по'
ru1 Russian rand 2024-12-04 20:31:20 797 Первая редакция (опубликовано)