OJ Board

I think it is very unfair to rejudge this type of problem.
Now you can see only Neal Zean can get the output completely correct.
It is very hard to get the exactly output for all the test cases. (I tried both of double and long double)
Because the data are floating-point numbers, we can just avoid ...

My ac code gets same output for your input

the source of these problems is same.
some of them have special judge.
but I think some correct output can't be accepted again.
except the judge solution itself.
hope you can fix it.

thanks very much,sohel!I will try it soon.
it looks like G,I got it in 0.060s.

who can give me some hint for this problem?
I always got TLE by backtracking + combinatorics.
what is the optimal algorithm?

thanks very much!now I get ac immediately by debugging step by step.
Add one condition and mark the some unnecessary start points.
hope it can help anyone who got wa for this problem!
thanks again,rio!

Anyone who got ac can give some tricky test cases?
I used bsearch and greedy,but got wa.
thanks in advance.


10

42 8 -3666 -3500
9169 5724 1478 -642
-3038 -5536 -4295 -1855
-6719 6827 -39 -9509
-7005 1942 -5173 -4564
2391 4604 -6098 -9847
-9708 2382 7421 8716
9718 9895 -4553 -8274
4771 1538 -8131 ...

For Problem 216 and 10750.
Accepted in the old system,but always got WA in the new.
I think the precision may be the main reason for P216.
But for 10750,my two versions O(n*n) and O(n*logn) both got WA.
hope you can check the test data and judgement,also with 10752
And 11087,always send to judge or ...

you are welcome!
you had ever helped me on some nasty problems.
I am testing and optimizing my ac codes for the problems I solved,found the precision is the main trouble in the new system.
no one can get ac on P216 now,it is an easy backtracking problem.
I think it must be a classical sample.

Leonid,have you got my code?
btw,my two version for P10751,n^2 and nlogn both got wa.
I think there must be something wrong in the special judge for this problem and P216

do you get the correct formula?
if your code passed jan's data,you should get AC.
the formula is quite simple and trivial.
if you need,I can send you my ac code by pm or email.

I got ac on P10751 when I print 4 digits after the decimal point.
I met precision problem several times in the new version.
I did the same for P10750,but wa...
btw,my 10750 code got ac in the old system.

my code output 20 for your last case.
others are same.
hope it helps

thank you,Rio!

I got it.I know this way to reduce the memory space.
But how to optimize the algorithm?
I think it only can save the memory in the dp.
but the graph array still needs 2000*2000.