|
|
| Показать все ветки Спрятать все ветки Показать все сообщения Спрятать все сообщения | | Hints | Denis Koshman | 1451. Басня о ларьке | 7 апр 2018 01:49 | 3 | Hints Denis Koshman 14 авг 2008 21:59 EPS=1e-8 precision for =0 comparisons is ok. Output 10 digits after decimal point is ok (6 digits gave WA41).
The point to find is called Fermat point, check at mathworld.wolfram.com Yea, Fermat point.... or simple ternary search :) Outputting 6 decimal points works just fine.
There is no need for epsilon tweaking: it can be solved with fractions (Python 3) except one Decimal (getcontext().prec=10) needed for sqrt(3). | | WA #2 | YuFeng | 1451. Басня о ларьке | 17 июл 2017 19:44 | 2 | WA #2 YuFeng 17 июл 2017 19:41 I don't know the test
I know exactly
Because I have verified it myself
That ternary search gives a solution
Even C float type is enough
And something about 200 iterations in each loop | | No subject | YuFeng | 1451. Басня о ларьке | 17 июл 2017 19:38 | 1 |
Edited by author 17.07.2017 19:41 | | all points on line | dibrov.bor[SumySU] | 1451. Басня о ларьке | 14 авг 2008 18:52 | 2 | what about case when all three points are on line
I think it's impossible to find required position as proper as we can't draw circle throw three points on one line Why do you think about the circle? The answer is obvious in such case, and it's the central point. | | WA1? | Samsonov Alex [USU] | 1451. Басня о ларьке | 2 апр 2007 01:12 | 2 | WA1? Samsonov Alex [USU] 2 апр 2007 01:00 Interesting. My solution gives correct answer on the sample test, but gets WA1 on Timus. Is the first test sample or not? Or, probably, there are some big differences between MSVC and Intel C++ Compiler? Re: WA1? Samsonov Alex [USU] 2 апр 2007 01:12 | | Changes in the problem (+) | Dmitry 'Diman_YES' Kovalioff | 1451. Басня о ларьке | 20 янв 2007 09:39 | 2 | Some tests were added. The checker precision was decreased. Look the statement for some changes in "Output" paragraph.
The problem has been rejudged. Some authors lost AC, and some get.
Many thanks to Fyodor Menshikov for new tests and help during the investigation.
Edited by author 19.01.2007 17:43 After changes especially after diminishing of precision
problem became absolutely simple and solvable numerically
by strict search on sequence of grids in double type
Last variant was very interesting due 1.e-8 |
|
|
|
