OJ Board

use cycle detection algorithm like http://en.wikipedia.org/wiki/Floyd_cycl ... _algorithm

Hint: split the vectors into smaller ones so that you search and removal time reduces.

You can improve the algorithm further. You don't need to n iterations to find what to print.

Hint: skip the numbers in block

-pvncad

I did the same thing. kept the partial sums of i^k S and
computed fF(k, a, b) using binomial expansion. Still I am getting WA?
it was running fine for sample inputs and all of my own testcases.

Could someone provide some inputs ?

andmej wrote:

What I miss the most is the e-mail received every time you submitted a problem.

It was very useful feature. I would love if we have it back.

[/quote]

I am getting WA. Any problem with following code ?


long long N = 0, n, res, nn;

while (1) {
scanf("%lld", &N);

if (N == 0) return 0;

if (N % 2 == 0) N --;

n = floor(sqrt(N)) - 1;

res = (n *( n + 1)) / 2;

if ( res % 3 == 0) {
res = (res / 3 ) * ( 2 * n + 1);
}
else {
res ...

Will try to recode and see how it goes. Thanks for ur help

i don't think that is a problem. num[j] represents the j-length digit starting from i + 1th position.

yup all r long long

I too felt the same but got so many WAs.

Can you anything wrong with this code ?

Actually, mine is also O(sqrt(2 *n))

I have solved this using O(sqrt(n)). emotional blind, any clue on o(1) solution ?

I tried to solve this using DP. But was getting WA.

Can someone give me tricky i/o ?

Thanks

9799

thanks. Now, I know what I missed out.