AdvancerMan's blog

By AdvancerMan, 5 years ago, translation, In English

1263A - Сладкая задача

Idea: MikeMirzayanov

Tutorial
Solution (MikeMirzayanov)

1263B - ПИН-коды

Idea: Stepavly

Tutorial
Solution (Stepavly)

1263C - Вы все уже победители!

Idea: unreal.eugene

Tutorial
Solution (unreal.eugene)

1263D - Секретные пароли

Idea: Stepavly

Tutorial
Solution (Stepavly)

1263E - Редактор

Idea: Supermagzzz

Tutorial
Solution (Supermagzzz)

1263F - Экономические проблемы

Idea: AdvancerMan

Tutorial
Solution (Rox)
  • Vote: I like it
  • +58
  • Vote: I do not like it

| Write comment?
»
5 years ago, # |
Rev. 2   Vote: I like it 0 Vote: I do not like it

Problem C can be done in O(√n) , I think. I just have a for loop 1->sqrt(n) and add 2 numbers to the list and got accepted :D

  • »
    »
    5 years ago, # ^ |
      Vote: I like it 0 Vote: I do not like it

    I think I have the same idea with you

  • »
    »
    5 years ago, # ^ |
    Rev. 2   Vote: I like it -7 Vote: I do not like it

    Edit: There was no need to use set or sort the vector. I was wrong.

    You need to either use a set, or if you used a vector then you need to sort them. So the factor of log(N) comes from there.

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

      For each $$$i \leq \sqrt{n}$$$, I calculate $$$h= \lfloor \dfrac{n}{\lfloor \frac{n}{i} \rfloor} \rfloor$$$. If $$$i=h$$$, I push back $$$i$$$ into a vector. If $$$t=\lfloor \frac{n}{i} \rfloor \neq i$$$, I also push back $$$t$$$ into another vector.

      Rearrange the order of printing elements from both vectors and you are good to go.

  • »
    »
    5 years ago, # ^ |
      Vote: I like it 0 Vote: I do not like it

    You add numbers to a set. Adding to the set is log(n). So your solution has O(sqrt(n)log(n))

»
5 years ago, # |
  Vote: I like it -8 Vote: I do not like it

For Problem B a similar solution would be to consider a map for storing all the found pins while iterating through the list. Once it finds a pin that already exists in the map, randomly choose any position in the pin to replace it randomly with another digit. This repeats until we get a new 4 digit number. And this counts as a step.

Finally, print the step count and the new pins.

https://mirror.codeforces.com/contest/1263/submission/66044121

»
5 years ago, # |
  Vote: I like it +15 Vote: I do not like it

C can also be solved using the fact that $$$j = \left\lfloor \dfrac{n}{\left\lfloor\dfrac{n}{i}\right\rfloor }\right\rfloor$$$ is the largest $$$j$$$ such that $$$\left\lfloor\dfrac{n}{i}\right\rfloor = \left\lfloor\dfrac{n}{j}\right\rfloor$$$.
65967376

»
5 years ago, # |
  Vote: I like it 0 Vote: I do not like it

Can anybody explain me logic of problem E. How it is done using segment tree?

  • »
    »
    5 years ago, # ^ |
      Vote: I like it +20 Vote: I do not like it

    Suppose that '(' is 1 and ')' is -1, and others are 0. If the string is a correct text, these conditions will be satisfied:
    - The sum of any prefix should not less than 0 because that means the number of ')' is greater than '(' in this prefix.
    - The sum of the whole string should be 0 because the number of '(' and ')' should be equal.
    If the two conditions are not satisfied, the answer is -1.
    Otherwise, the answer is equal to the maximum depth of these brackets. It is equal to the maximum sum prefix because the sum of a prefix means how much '(' do not match with a ')' in the prefix.
    So we can use a segment tree to keep the sum of all prefixes and get the maximum and minimum sum prefix.
    My code: 66065668

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

      Can you explain your solution a little bit. What is the use of tag.

      • »
        »
        »
        »
        5 years ago, # ^ |
          Vote: I like it 0 Vote: I do not like it

        I use a segment tree to keep prefix sum array, so when I modify the $$$i$$$-th element in the original array, I need to update sums of prefixes that contain the $$$i$$$-th element. They are a continuous range $$$[i, n-1]$$$ in the prefix sum array, so we need to use lazy tag to do range updates. If the tag of a node is $$$t$$$, that means every element in the range should plus $$$t$$$, and the maximum and minimum value of the node should plus $$$t$$$, too.

        • »
          »
          »
          »
          »
          4 years ago, # ^ |
          Rev. 3   Vote: I like it 0 Vote: I do not like it

          I did it using segment tree but getting TLE on the 8th test case. Whats the problem with my code.

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

            mindsweeper You have used long long int numbers in your code. I think that is the reason. Use int and get AC :)

    • »
      »
      »
      5 years ago, # ^ |
        Vote: I like it 0 Vote: I do not like it

      What's the time complexity? Is it n^2*log(n) ?

      • »
        »
        »
        »
        5 years ago, # ^ |
          Vote: I like it 0 Vote: I do not like it

        No, the time complexity is $$$O(n \log n)$$$.
        The time complexity of doing a range update or range minimum/maximum query is $$$O(\log n)$$$. In my code, when I process a command, I do range update at most twice and range minimum/maximum query at most 3 times, so the time complexity of processing a command is $$$O(\log n)$$$. There is $$$n$$$ commands, so the time complexity is $$$O(n \log n)$$$.

»
5 years ago, # |
  Vote: I like it +3 Vote: I do not like it

How to solve problem E using lazy segment tree? :)

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

A different O(n) approach for E:

Maintain a prefix sum and store maximum and minimun in it.

Be lazy when the operation is L\R or when #( != #) (number of brackets).

when #( = #), update the prefix sum just in the range the cursor touched, amortize time is O(1) since L,R operations paid for each update in this range.

To maintain maximum and minimum store an histogram of prefixes.

solution: 66080615

  • »
    »
    5 years ago, # ^ |
      Vote: I like it +3 Vote: I do not like it

    i was thinking in this way, but then i've remembered that i'm dummy and came with obvious segment tree solution

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

Sweet candies problem If we are given 8 2 8 candies, there is no possible way that we can eat it for 9 days. Can you tell me what is wrong in my approach?

  • »
    »
    5 years ago, # ^ |
      Vote: I like it +5 Vote: I do not like it

    It is possible that we can eat candies for 9 days. One scenario for example would be that we eat 7 candies from biggest and medium piles i.e. 8 and 8 here. This can be done for 7 days. Then we are left (1 2 1) candies. Then eat such that the pile is (0 1 1) on the 8th day and on 9th day, it would be (0 0 0).

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

Can anybody explain the Segment tree solution of E problem a little elaborately, I have read solutions of many people but couldn't decipher the logic of what they are trying to do. I know these things that the minimum sum prefix has to be greater than -1, total sum has to be 0 and maximum sum prefix would the answer, but how exactly are we computing this using a segment tree?

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

    Suppose you know that the total length of everything is n. Let each '(' to be a +1, each ')' to be a -1, and each regular character to be a 0.

    So for example if I have string (((a))), we get [1, 2, 3, 3, 2, 1, 0]. Notice that if we drop below 0 then we have an invalid string since there are too many ')'. The maximum amount of nesting is what we want to query, and that's just the maximum value on the segment.

    When we make changes we can just perform updates on segment. For example if I change from 'a' -> '(' at index i, then I just add 1 for all elements from [i...n], if I change from ')' to '(', then I have to add 2 for all elements [i...n]. Really we have range sum updates and range max/min queries which a straightforward use of segment tree

    • »
      »
      »
      5 years ago, # ^ |
      Rev. 3   Vote: I like it 0 Vote: I do not like it

      Thank you so much, this explanation is crystal clear. But I wanted to ask you this that changing a a to ( is a point update so can't we do that point update and calculate the max prefix and min prefix sum using some other way? I thought of a way that we store totalSum, MaxPrefixSum and MinPrefixSum for each node, now for every parent thw updation would be something like parent.MaxPrefixSum = max(Left.MaxPrefixSum, Left.TotalSum+Right.MaxPrefixSum) And similarly for MinPrefixSum TotalSum would be updated directly.

»
5 years ago, # |
Rev. 4   Vote: I like it +8 Vote: I do not like it

Challenge Problem F completed:

The complexity of my code is linear. I only use algorithms like dfs, counting sort, monotonic stack.

My solution

»
5 years ago, # |
  Vote: I like it +8 Vote: I do not like it

(In the editorial of problem A)Is the line correct? "Then we make equal the piles g, b by eating g−b from the piles r and b".

I think it would be-"Then we make equal the piles g, b by eating g−b from the piles r and g".

  • »
    »
    5 years ago, # ^ |
      Vote: I like it 0 Vote: I do not like it

    You are right, thanks for pointing it out! The tutorial for A will be updated soon.

»
5 years ago, # |
  Vote: I like it +1 Vote: I do not like it

How to solve problem E with segment tree??? plz explain for me. thanks..

  • »
    »
    5 years ago, # ^ |
      Vote: I like it 0 Vote: I do not like it

    Lets assume '(' = 1 and ')' = -1 and any other character is equal to 0. For each segment store the following things:-
    1. Number of opening brackets.
    2. Number of closing brackets.
    3. Maximum prefix sum and minimum prefix sum.

    For each iteration update the tree. The string is balanced if minimum prefix sum is positive or zero and number of opening brackets == number of closing brackets. Number of different colors (final answer) is maximum prefix sum. You can have a look on my submission.66222540

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

Tutorial for problem has a typo I guess: Then we make equal the piles g, b by eating g−b from the piles r and b.

Here it should be piles r and g I think. Correct me if I am wrong.

»
5 years ago, # |
  Vote: I like it 0 Vote: I do not like it

Can someone explain how to solve Div2-F. I am not able to understand the editorial. P.S- What is " the segment [l,r]", I am not able to understand it. Thanks in advance.

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

Can anyone please tell me where my approach of Problem B is wrong .

https://mirror.codeforces.com/contest/1263/submission/66242150

It's even showing the wrong answer on test case 1, which i can't seems to understand.

EDIT: Found my mistake !!

»
5 years ago, # |
  Vote: I like it 0 Vote: I do not like it

Can you tell me why for the input 8 9 10

output is 13, it should be 16.

  • »
    »
    5 years ago, # ^ |
      Vote: I like it 0 Vote: I do not like it

    Tanya can't eat for $$$16$$$ days straight because the sum of piles' sizes is reducing by $$$2$$$ every day so she can't eat for more than $$$\left\lfloor \frac{8 + 9 + 10}{2} \right\rfloor = 13$$$ days.

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

My submission is 66433037 and it's completely linear but it's giving me TLE for problem E?? Can anyone help me fix this? I've used a Stack of an object which is 4 generics, and all of my actions are push and pop and writing to an array; not really sure where I could clean this up; is there some Java structure I'm using which is particularly slow?

»
5 years ago, # |
  Vote: I like it 0 Vote: I do not like it

For D, 4 l k al ak Answer is 1. Can someone tell me how l and k are equivalent?

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

AdvancerMan isn't the recurrence given in F incorrect?

If you're using 0-based indexing it should be: $$$dp_i = \max\limits_{0 \le j < i}(dp_{j}+ max(upperCost_{j, i-1}, lowerCost_{j, i-1}))$$$, where $$$dp_0 = 0$$$.

»
5 years ago, # |
  Vote: I like it 0 Vote: I do not like it

I implemented it in the same way but get WA verdict. Someone please help 67183675

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

Hello AdvancerMan. I found that maybe the test cases for problem D are still not strong enough. For example:

3
cd
bc
ad

This case will output 1 for my submission 73597744, output 2 for my submission 73603502 , but both of the two submissions are accepted now.

I think the right answer should be 1. Do I have any misunderstanding?

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

    Hello! I've checked your solutions and you don't have any misunderstanding. Indeed answer for the given test case is 1.

»
4 years ago, # |
  Vote: I like it 0 Vote: I do not like it

Hi, AdvancerMan

I get denial of judgement when upsolving problem F.

I even tried other's accepted code and it still crashes, please look into it.

https://mirror.codeforces.com/contest/1263/submission/84879747

The error message goes like this:

Test: #5, time: 0 ms., memory: 0 KB, exit code: 0, checker exit code: 0, verdict: CRASHED Checker Log

Generator is not determinate [the verification run produces different output, cmd =gen 1000 3 1000 2 1000 1 0 200 3 1000 2 1000 1 0 150], [16990 bytes, '1000

1244

1 970 1 1239 533 226 912 1 824 1 576 9... 1009 75 419 864 541 906 735 335 380 496 1097

', sha1=1ba04e177afba056927a2855f07e8327356b8113] vs. [16999 bytes, '1000

1244

1 970 1 1239 533 226 912 1 824 1 576 9...332 1164 609 569 742 895 1049 123 974 690 852

', sha1=40211f2f08ea2d3d7248b9d2960c5dd4cbd04154].

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

Solution to problem D by DSU : 263748631

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

how to solve problem c by binary search and mid in the middle?