OJ Board

david
Most probably - yes. I think that the proof that R is indeed transitive will solve everything, once internals of this proof are given.

Another noted thing which makes me feeling "unsafe" is that "unique resulting string means unique ordering of pieces". 123456 can be 123+456 or 12+34+56 ...

david
I understand your way of proving it via "we have the only one possible answer", and then applying this property it to the original data.

What "irritates" me a little is that whenever I tried to solve some other problems this way, I always found a flaw in my decisions like A -> B, and B -> A ...

lord_burgos
Please correct your forumlae... I don't see why 'S' stopped being an exponent

david
223++22 > 22++223
but
22 4 223 > 223 4 22 (if '4' is placed in between 'a' and 'b'), so this is not true for arbitrarily placed pairs.

Porbably it still can be proven via adjacent swaps, but it is not obvious for me how to prove that bubble sort local maximums will yield global maximum ...

My answer is 4-3-4-3-3-2

My AC solution performed edge-split (described earlier in this topic) on every edge. All weaker efforts failed, although they were more time-effective.

My solution outputs
2
2
run-time-error (duplicate edge)

Without that checking it outputs 2-2-1-2-2-5

This should be TL, but not WA... Check your loops. Is it O(N^9)?

You have to find 2 paths from S to T, such that they don't share any edge (but may share some nodes). Sum of their lengths must be minimal possible.

Adrian,

My thoughts told that it is better to give wins to teams who are above us to ensure most zeroes behind. Kinda greedy method. Then distribute draws as much as possible. Give as much wins to yourself as possible, and so, and so... I tried to write this 4 times, and each time threw everything ...

As you may see in my previous message, try to guess the formula. It's much easier than all simulation code behind it... It appears like sin(asin(sqrt(x^2))) = (?) = x. Not that easy formula, but close.

I used 2D ternary search (nestes 1D's).

With O(n*4^s) DP I got TL during the contest. It's better to index back and forth all 3^s states into 4^s map.

Age 21 (almost 22), registered AFAIK in 1999 (too lazy to check my INFO ). I usually solve about 200 problems in 2-6 weeks, then I understand that it's really time to get back to work, and evantually 6-12 months gap follows.

P.S: my ID is prime (4153)

ihi,

Try to find some easy to calculate lower and upper bounds for result and check if your real algorithm's result fits into this range.