Блог пользователя maroonrk

Автор maroonrk, история, 21 месяц назад, По-английски

We will hold AtCoder Regular Contest 145.

The point values will be 400-500-600-700-800-1200.

We are looking forward to your participation!

  • Проголосовать: нравится
  • +155
  • Проголосовать: не нравится

»
21 месяц назад, # |
  Проголосовать: нравится +32 Проголосовать: не нравится

Looking forward to ARC !!

»
21 месяц назад, # |
  Проголосовать: нравится +151 Проголосовать: не нравится

As the writer, good luck and enjoy yourself! :D

»
21 месяц назад, # |
  Проголосовать: нравится +1 Проголосовать: не нравится

What's the approximate rating of ARC's problems B & C according to CodeForces rating distribution?

»
21 месяц назад, # |
  Проголосовать: нравится +19 Проголосовать: не нравится

math and counting round

»
21 месяц назад, # |
  Проголосовать: нравится +6 Проголосовать: не нравится

I really suck at counting, but today I couldn't figure out what the 16 different permutations were in C, the most I found was 12 :P

Can anyone list those out please?

  • »
    »
    21 месяц назад, # ^ |
      Проголосовать: нравится +22 Проголосовать: не нравится
    1 2 3 4
    2 1 3 4
    1 2 4 3
    2 1 4 3
    3 4 1 2
    4 3 1 2
    3 4 2 1
    4 3 2 1
    1 3 2 4
    2 3 1 4
    1 4 2 3
    2 4 1 3
    3 1 4 2
    4 1 3 2
    4 2 3 1
    3 2 4 1
    
  • »
    »
    21 месяц назад, # ^ |
      Проголосовать: нравится +1 Проголосовать: не нравится
    Code that prints
»
21 месяц назад, # |
  Проголосовать: нравится +1 Проголосовать: не нравится

Could someone point out what case am I missing in my submission for B?

Submission : https://atcoder.jp/contests/arc145/submissions/33632229

»
21 месяц назад, # |
  Проголосовать: нравится 0 Проголосовать: не нравится

My first AtCoder contest. The problems were good, it was interesting for me. Thanks !

»
21 месяц назад, # |
  Проголосовать: нравится +65 Проголосовать: не нравится

D is just cantor set...

»
21 месяц назад, # |
  Проголосовать: нравится +10 Проголосовать: не нравится

As a contestant, I was not a big fan of A,B,D. Those problems' ideas are not hard to get, but have very annoying corner cases...

»
21 месяц назад, # |
  Проголосовать: нравится 0 Проголосовать: не нравится

Problem D is magic, but I want to know, why the set $$${1,2,4,7,11,...}$$$, which the differential sequence is $$${1,2,3,4,...}$$$ is not the optimal one.

  • »
    »
    21 месяц назад, # ^ |
      Проголосовать: нравится +19 Проголосовать: не нравится

    The set $$$1,2,4,7,11,\dots$$$ has an $$$\mathrm{O}(n^2)$$$ value on the $$$n$$$-th element, while the optimal set has $$$\mathrm{O}(n^{\log_2 3}) \approx \mathrm{O}(n^{1.585})$$$ value on the $$$n$$$-th one ,which is enough to meet the constraint of $$$[-10^7,10^7]$$$ .

    • »
      »
      »
      21 месяц назад, # ^ |
        Проголосовать: нравится 0 Проголосовать: не нравится

      U R right.I tried that,and the number can be as large as 4e7

  • »
    »
    21 месяц назад, # ^ |
      Проголосовать: нравится +30 Проголосовать: не нравится

    That doesn't even work, $$$1+7=2\cdot 4$$$

»
21 месяц назад, # |
  Проголосовать: нравится +16 Проголосовать: не нравится

Sequence of pC can be found here.

»
21 месяц назад, # |
  Проголосовать: нравится +24 Проголосовать: не нравится

My friend told me the problem C is in the OEIS, what do you think of it?

»
21 месяц назад, # |
  Проголосовать: нравится 0 Проголосовать: не нравится

void jaiishriRam(){ ll n; cin>>n; string s; cin>>s; if(n==2) { if(s=="BA" || s=="AB") { cout<<"NO"<<endl; return; } } if(s[0]=='A' && s[n-1]=='B') { cout<<"NO"<<endl; return; }

cout<<"YES"<<endl;

} could anyone help me to know what thing I did wrong in A above is the code

»
21 месяц назад, # |
  Проголосовать: нравится +16 Проголосовать: не нравится

Can anyone plz tell me whats wrong here in my submission for problem B

  • »
    »
    21 месяц назад, # ^ |
    Rev. 2   Проголосовать: нравится +1 Проголосовать: не нравится

    Consider the test case :- 6 5 4 ,
    Your O/P — 4
    Correct O/P — 2
    for value (5 and 6 only Alice will win )

    • »
      »
      »
      21 месяц назад, # ^ |
        Проголосовать: нравится 0 Проголосовать: не нравится

      Corrected that mistake but still getting WA on 3 test cases ;-; Here is my submission

      • »
        »
        »
        »
        21 месяц назад, # ^ |
        Rev. 2   Проголосовать: нравится +1 Проголосовать: не нравится

        I'm not sure if binary search is a good idea here. You can try 30 10 3 correct ans: 7 your code gives: 9

      • »
        »
        »
        »
        21 месяц назад, # ^ |
          Проголосовать: нравится +1 Проголосовать: не нравится

        Consider this test case :- 8 4 3
        Your O/P — 6
        Correct O/P — 4
        Aice wins game (4,5,6 and 8)
        Error is in line 58 loop that you are multiplying . They will depend on remainder of (n%a).

»
21 месяц назад, # |
  Проголосовать: нравится +8 Проголосовать: не нравится

Can anyone please tell me whats wrong here in my submission for problem D?

  • »
    »
    21 месяц назад, # ^ |
      Проголосовать: нравится +16 Проголосовать: не нравится

    I don't know how your code works, but it fails at the input 4 2

    • »
      »
      »
      21 месяц назад, # ^ |
        Проголосовать: нравится +8 Проголосовать: не нравится

      thx.

    • »
      »
      »
      21 месяц назад, # ^ |
        Проголосовать: нравится +8 Проголосовать: не нравится

      It works in such way:

      $$$\pi(n)\sim \frac n{\ln n}$$$

      So the dfs is expected to stop in very low depth.

      p.s. What about this one? It could pass 4 2.

      • »
        »
        »
        »
        21 месяц назад, # ^ |
          Проголосовать: нравится 0 Проголосовать: не нравится

        It fails at 10000 0.

        The output has these three integers:

        -9999991 -9999937 -9999883
        
        • »
          »
          »
          »
          »
          21 месяц назад, # ^ |
            Проголосовать: нравится +8 Проголосовать: не нравится

          Sorry, but it got 9999991 -9999991 9999973 -9999973 9999971 ... here.

          Maybe there's some Undefined Behavior in my code?

          • »
            »
            »
            »
            »
            »
            21 месяц назад, # ^ |
              Проголосовать: нравится +8 Проголосовать: не нравится

            It is not a UB, because the output is a set, and these three numbers are in it, which is against the constraint.

            Actually, they are not the first three numbers of the output on my computer as well.

      • »
        »
        »
        »
        21 месяц назад, # ^ |
          Проголосовать: нравится +27 Проголосовать: не нравится

        Are you assuming primes don't have arithmetic sequences of length 3? Because that is not true, for example: $$$101, 107, 113$$$.

»
21 месяц назад, # |
  Проголосовать: нравится +40 Проголосовать: не нравится

C could be solved by writing a brute force and searching the sequence on OEIS. Maybe those problems should be avoided by for example having problems parametrized by some more values, or making sure it isn't easy to find on the internet. It makes the contest experience less fun, if the answer is easily found on the internet. Problems A and B were fine. Problem D also suffered from this issue.

»
21 месяц назад, # |
  Проголосовать: нравится +18 Проголосовать: не нравится

C is actually A152029(OEIS).

»
21 месяц назад, # |
Rev. 3   Проголосовать: нравится 0 Проголосовать: не нравится

Why does this code give WA for problem A?

lng dp[3][SZ],n;string s;
lng solve(lng i,lng x)
{
 if(i==n-i){return 1;}
 if(i>n-i){return (x==0);}
 if(dp[x][i]!=-1){return dp[x][i];}
 lng j=(n-1)-i;
 char a=s[i],b=s[j];
 if(x==1){a='B';}
 if(x==2){b='A';}
 if(a==b)
 {
  lng z=solve(i+1,0);
  if(a=='A'){z=max(solve(i+1,1),z);}
  else{z=max(solve(i+1,2),z);}
  return dp[x][i]=z;
 }
 if(j-i==1){return 0;}
 if(a=='A'){return 0;}
 return dp[x][i]=max(solve(i+1,1),solve(i+1,2));
}
int main()
{
 ios_base::sync_with_stdio(0);cin.tie(0);cout.tie(0);
 cin >> n >> s;
 for(lng i=0;i<3;i++){for(lng j=0;j<SZ;j++){dp[i][j]=-1;}}
 if(solve(0,0)==1){cout << "Yes";rtr;}
 cout << "No";rtr;
 return 0;
}

Here lng=long long, rtr=return 0, SZ=2e5+7

»
21 месяц назад, # |
  Проголосовать: нравится 0 Проголосовать: не нравится

This was my first contest on AtCoder, Can someone tell me, when do the editorials be out ?

»
21 месяц назад, # |
  Проголосовать: нравится +8 Проголосовать: не нравится

nok0 My submission for A get AC but can be hacked with n=4 AAAA My submission

»
21 месяц назад, # |
Rev. 2   Проголосовать: нравится +3 Проголосовать: не нравится

In problem A:

Why is "AB" a valid palindrome or can be changed to a palindrome according to editorial and tests?

»
21 месяц назад, # |
  Проголосовать: нравится +8 Проголосовать: не нравится

for D you can just use brute force to get a set meet the last restriction,and change the largest/smallest number,then add x to all numbers to get sum=M

base 3 solution is beautiful but not necessary

»
21 месяц назад, # |
  Проголосовать: нравится +117 Проголосовать: не нравится

I have different approaches for problem D and E:

D:

Ingore the limit of sum for now. Let $$$f(n)$$$ be the set of integers, one can find that

$$$ f(1)=\{0\} \\ f(2n)= f(n) \cup (f(n)+2\max f(n)+ {\color{red}1}) $$$

is good. Now consider the limit of sum. One can find that if we change the $$$\color{red}1$$$ into $$$2$$$, and add $$$1$$$ to some greatest elements in $$$f(n)$$$, it is still good.

E:

We can change $$$A_j$$$ to $$$A_i\oplus A_j$$$ ($$$i<j$$$) by performing operations $$$j,i+2,i+3,\ldots,j$$$ in order.

Now we change $$$A_i$$$ into $$$B_i$$$ from right to left, taking the lexicological largest basis each time. It turns out that the number of operations is OK. (It can take up to $$$N-1$$$ operations to change a single number, but each operation will move some basis vectors to the right)

  • »
    »
    21 месяц назад, # ^ |
      Проголосовать: нравится +100 Проголосовать: не нравится

    Your code on E submitted during contest can be hacked using the data generated by the code below.

    #include<bits/stdc++.h>
    #define rep(i,a,n) for (int i=(a);i<=(n);i++)
    
    int main() {
    	int n = 1000;
    	printf("%d\n", n);
    	rep(i, 0, 59) printf("%lld ", 1ll << i);
    	rep(i, 1, n - 60) printf("%lld ", (1ll << 60) - 1);
    	printf("\n");
    	rep(i, 0, 59) printf("%lld ", 1ll << i);
    	rep(i, 1, n - 60) printf("%lld ", (1ll << 60) - 1 - (i & 1));
    	return 0;
    }
    
    • »
      »
      »
      21 месяц назад, # ^ |
        Проголосовать: нравится +6 Проголосовать: не нравится

      Thanks for pointing this out! I tried to add some randomization and it worked pretty good.

»
21 месяц назад, # |
Rev. 2   Проголосовать: нравится +39 Проголосовать: не нравится

My randomized solution for E:

If the answer is Yes, we first randomly manipulate the sequence for a small number of times, and then operate on index $$$n$$$ more than $$$n$$$ times repeatedly. Then we can (roughly) assume that $$$a_n \oplus b_n$$$ is a linear combination of $$$a_{n-r+1 \cdots a_{n-1}}$$$, where $$$t$$$ is about $$$60$$$.

We can then solve $$$a_n \oplus b_n = \bigoplus_{j \in S} a_j$$$ with Gaussian elimination. If $$$a_{n-1} \not \in S$$$, operate on $$$n$$$ so that $$$a_{n-1} \in S$$$.

Then we operate on the second minimum element $$$t$$$ in $$$S$$$. Then $$$\min S$$$ is removed from $$$S$$$ and all indices between $$$\min S$$$ and $$$t$$$ are added. Repeat this process until $$$S = {a_{n-1}}$$$ and operate on $$$n$$$. Do this for $$$n - 1 \dots 2$$$ and we are done. This solution uses ~60000 operations in randomly generated inputs.

implementation

UPD: More than 1000 random operations at first are needed to guarantee randomness.

»
21 месяц назад, # |
  Проголосовать: нравится 0 Проголосовать: не нравится

Can somebody explain here problem D, how exactly are we constructing the required set after we have figured out the base 3 condition where each digit should be 0/1.

»
21 месяц назад, # |
  Проголосовать: нравится +10 Проголосовать: не нравится

F is some wild generalization of IMO 1995 P6. I like it

»
21 месяц назад, # |
  Проголосовать: нравится +5 Проголосовать: не нравится

Such a magical problem D!

Base-3 representation is such a trick that could perfectly beat the rule of $$$2y \neq x + z$$$.

Next, by first keeping the rightmost digit of all n integers as zero and then change some of them to 1, so that $$$|m - sum| \% n = 0$$$. At the same time, all n integers are still distinct and satisfy the "good set" rule.

Finally, add $$$(m - sum) / n$$$ to each integer so that $$$sum = m$$$, and at the same time, adding the same integer will not lead to any $$$2y = x + z$$$, although they may not belong to the "good set" anymore.

Thanks to the problem writer for coming up with such a clever and unbelievable construction problem!

»
21 месяц назад, # |
  Проголосовать: нравится 0 Проголосовать: не нравится

This round is good, arc should contain magical counting and constructing problems like this.

»
21 месяц назад, # |
Rev. 4   Проголосовать: нравится 0 Проголосовать: не нравится

A different way to think for D.

We can construct as follow (P is a set,initial $$$P={1},len=1$$$): $$$P\to P\cup (P+(2len-1)),len\to 3len-1$$$

$$$P+a$$$ means add $$$a$$$ to all elements in $$$P$$$.

And it can prove that $$$y−x\not=z−y$$$ for every triple $$$x,y,z (x<y<z)$$$ of distinct elements in $$$P$$$.

(Which is similar to the official editorial)

and we can construct the set with $$$n+2$$$ elements, and enumerate which two to delete. It can solve the problem.

  • »
    »
    21 месяц назад, # ^ |
      Проголосовать: нравится 0 Проголосовать: не нравится

    Could you please talk more about why your construction will not lead to 2y=x+z? And, how to construct these n+2 elements, and how to determine which two to delete?

    Thank you so much for providing such a clever idea.

»
20 месяцев назад, # |
  Проголосовать: нравится +1 Проголосовать: не нравится

Sorry for asking it late,

Can any one please the explain this line of editorial :

Here, if the elements in P are scanned from left to right, there must always be more left elements than right elements. The number of such arrangements corresponds to that of parenthesis sequences.

What I understood is that, lets say we have (1,2),(3,4),(5,6),(7,8) as corresponding (ai and bi).
Permutation : (1,2,4,3,5,6,7,8) have more right elements(2) than left elements(1) in first 3 elements (1,2,4) but still it can be divided it into two sequences A(1,3,5,7) and B(2,4,6,8), but in editorial it is told that it should not be possible ( as per I understood ) .
I know I have understood it wrong.
Anyone who did this problem or got the editorial please explain the line in blue

Editorial link : Link

»
20 месяцев назад, # |
  Проголосовать: нравится 0 Проголосовать: не нравится

sorry for asking late. i'm confused about problem C on how/why the legal arrangement corresponds to parenthesis sequences. (that seems strange to me)

i know we are arranging some pairs like (1,2)(3,4)... , but in my opinion and based on some experiments and some comments like this. i claim an arrangement is legal if and only if there is no pair contains another pair.

for example, (1,2,3,4) legal because pair(1,2) and (3,4) doesn't contains the other. (1,3,2,4) is legal because two pair intersect not contain. (1,3,4,2) not legal because of containing.

however, if you see the pair as parentheses, (1,3,4,2) is legal because they can be mapped to "(())". that's a contradiction (do i misunderstand something?)

however again, when we consider the number of it, these two somehow are equal. here i take some example on n=3 to illustrate.

you see, if we take the view of parentheses, their correspondence(left parentheses and right parentheses) is really strange (at least differ from matching them greedily)

in general, i'm not negate the editorial, but i think i need some more clear intuition to understand this transform