AminAnvari's blog

By AminAnvari, 23 months ago, In English,

Hi :)
Some problem about powerful data structure and algorithm SQRT decomposition.
If you want to learn this algorithm click here.
And a big thank to mr_agha_seyed for helping.



220B - Little Elephant and Array
86D - Powerful array
13E - Holes
455D - Serega and Fun
375D - Tree and Queries
446C - DZY Loves Fibonacci Numbers
487D - Conveyor Belts
506D - Mr. Kitayuta's Colorful Graph
348C - Subset Sums
617E - XOR and Favorite Number
444C - DZY Loves Colors
398D - Instant Messanger
342E - Xenia and Tree

Maybe some of them can solve by other data structures like segment tree.

 
 
 
 
  • Vote: I like it  
  • +168
  • Vote: I do not like it  

»
23 months ago, # |
  Vote: I like it +5 Vote: I do not like it

cheers! :) :D

»
23 months ago, # |
  Vote: I like it +2 Vote: I do not like it

Would be nice to have SQRT decomposition tag for those problems. Now, I guess, all of them in data structures

http://codeforces.com/problemset/tags/data%20structures

»
23 months ago, # |
  Vote: I like it +6 Vote: I do not like it

timus 1846 also nice sqrt-decomp problem.

  • »
    »
    23 months ago, # ^ |
      Vote: I like it 0 Vote: I do not like it

    Can you, please, elaborate a bit?

    • »
      »
      »
      23 months ago, # ^ |
      Rev. 5   Vote: I like it +3 Vote: I do not like it

      I'm thinking of splitting the queries in chunks of sqrt(N) and processing them similar to the disjoint-set-union split as described on e-maxx.

      First of all, we will call a set a DS which can handle the insert and erase operations in O(1).

      Let active set and temp. set be 2 empty sets

      For every chunk of sqrt(N) queries:

      1. Go through all the remove operations (query type '-') and, if possible, remove those numbers from the active set of numbers, and add them in a temp. set. This loop will take O(sqrt(N)).

      2. Let currGcd be the gcd of all the keys in our active set. There are at most O(N) of them.

      3. Process the chunk again and take 2 cases:

      • The query asks to remove X. We will erase it from our temp. set.
      • The query asks to insert X. We will insert it in the temp. set.

      The answer for every query will be the gcd of all the keys in the temp. set and our currGcd. There are at most O(sqrt(N)) items in the temp. set, so this loop will take O(N).

      At the end of a chunk, move all the items from the temp set to the active set (and clear the temp set).

      The overall complexity is O(N * sqrt(N))

    • »
      »
      »
      23 months ago, # ^ |
      Rev. 2   Vote: I like it 0 Vote: I do not like it

      можно хранить наши элементы по блокам (gcd всех элементов в нем) + хранить мапу, где запоминаем, в каком блоке лежит число, для ответа пройдемся по блокам и пересчитаем gcd, для добавления/удаления пересчитаем gcd только в одном блоке код

  • »
    »
    10 months ago, # ^ |
      Vote: I like it 0 Vote: I do not like it

    It easily solves with segment tree, each leaf node stores x, non-leaf nodes gcd of its children.

»
23 months ago, # |
  Vote: I like it +11 Vote: I do not like it

Are Mo Algorithm and SQRT decomposition considered the same?

In SQRT decomposition we divide the array in buckets, solve the task in these buckets and answer each query in O() time (worst case). However, using Mo Algorithm the queries are sorted according to the bucket the left end of the queries falls into ...

I think both methods have the same time complexity, but are they considered the same?

  • »
    »
    23 months ago, # ^ |
      Vote: I like it +23 Vote: I do not like it

    SQRT decomposition combines several approaches and algorithms. Mo Algorithm is one of them.

  • »
    »
    23 months ago, # ^ |
      Vote: I like it -10 Vote: I do not like it

    SQRT Descomposition is a data structure, while Mo's Algorithm is more a technique.

    • »
      »
      »
      23 months ago, # ^ |
        Vote: I like it 0 Vote: I do not like it

      Not really much of a proof. SQRT-decomposition on queries isn't so much of a data structure, but is still usually meant to be called SQRT-decomp

»
22 months ago, # |
  Vote: I like it 0 Vote: I do not like it

I got stuck with the problem 446C — DZY Loves Fibonacci Numbers and here's my submissions

could someone help ?

thanks in advance

»
10 months ago, # |
  Vote: I like it 0 Vote: I do not like it

Can 13E - Holes solved using MO's algorithm? Or There are other kind of Sqrt-Decomposition? I have solved the first two in the list. But can't find an idea about this one.

  • »
    »
    10 months ago, # ^ |
      Vote: I like it 0 Vote: I do not like it

    You know what, there is this thing called editorial for (almost) every round which explains how to solve the problems.

    • »
      »
      »
      10 months ago, # ^ |
        Vote: I like it 0 Vote: I do not like it

      Thank you. I read the editorial before the comment. But their solution doesn't feel like the MO I learnt. And this problem is listed here. So I asked here if the solution is general, there might be some name or insights available that I want to know before solving it.

  • »
    »
    10 months ago, # ^ |
      Vote: I like it +8 Vote: I do not like it

    It is explained there 23:05

»
10 months ago, # |
  Vote: I like it 0 Vote: I do not like it

242E - XOR on Segment can also be solved with SQRT-decomposition. Although it's slower than the intended solution, it can still AC.

»
10 months ago, # |
  Vote: I like it 0 Vote: I do not like it

One more very nice SQRT-decomposition problem I encountered at the Innopolis Open Elimination round (2016-2017): http://codeforces.com/gym/101182/problem/E

»
5 months ago, # |
Rev. 2   Vote: I like it 0 Vote: I do not like it

Is it better to use const value for the sqrt (i.e 350) or to use sqrt function on n ?

»
4 months ago, # |
Rev. 2   Vote: I like it +3 Vote: I do not like it

Why i am getting TLE at 220B - Little Elephant and Array? My Submission 29417078 .Could anyone help me!

Edited: I have changed my solution from map to array.But this also gives me a TLE! 29417554

  • »
    »
    4 months ago, # ^ |
      Vote: I like it 0 Vote: I do not like it

    Maybe you have to create a struct for the query, add and remove method you can improve it because you have a lot process and you can put this in you cmp.

    bool cmp(const query& p,const query& q){
        if(p.l/nb != q.l/nb)
            return p.l/nb<q.l/nb;
        return ((p.r<q.r)^((p.l/nb)&1));
    }
    
»
4 months ago, # |
  Vote: I like it +5 Vote: I do not like it

https://www.codechef.com/AUG17/problems/HILLJUMP Another good problem on this topic.

»
3 months ago, # |
  Vote: I like it 0 Vote: I do not like it

It's great to see guys like you guide beginners like us ! Thank you so much !

»
2 months ago, # |
  Vote: I like it 0 Vote: I do not like it

446C DZY Loves Fibonacci Numbers gives TLE with sqrt decomposition. My code gives TLE although its complexity is O(N*sqrt(N))?? Any help would be appreciated.

  • »
    »
    2 months ago, # ^ |
      Vote: I like it +1 Vote: I do not like it

    I don't find your mistake, but i know easier way to solve it.

    See my code.

    Main idea : divide queries in sqrt(m) blocks.

    • »
      »
      »
      2 months ago, # ^ |
        Vote: I like it 0 Vote: I do not like it

      I saw your code. Your Idea is cool. But your code and mine have literally the same complexity.