OJ Board

My simple C++ implementation of Kosaraju's algorithm runs in 0.7 sec, so yours is probably slower because of language. What language do you use?

EDIT: oh, I see you code in C++. So it might be because of cin/cout instead of scanf/printf.

Thanks, Per
With your hint, I found exactly the same formula, without reading the spoiler

Could anyone tell me an idea how to solve this problem?

Thank you, I've got AC

Can anyone explain an idea how to solve this problem?

Thanks you all, I got AC, but only after setting precision of V to 1e-10 (1e-8 didn't work).

I used dynamic programming to find the best velocity for the x-th segment if I arrive with y crashes. When computing this, the corresponding function T(v) (input: speed on the x-th segment, output: expected time to finish the whole track, valid inputs: (0,M-y> ) has only one local minimum and this ...

Can anyone give the idea to solve this problem?
Is it true that function T(v) has only one local extremum?