Please subscribe to the official Codeforces channel in Telegram via the link https://t.me/codeforces_official. ×

### -is-this-fft-'s blog

By -is-this-fft-, history, 19 months ago,

I wanted to create this blog because there will probably not be really much discussion of solutions in the award ceremony (currently in progress). Also people who aren't finalists themselves can't see that anyway.

I'm interested to hear approaches of the higher-scoring finalists. (We, TÜ Randomized Algorithms did not do anything interesting :D)

A bit on the negative side: this seemed to be not be a 4-hour contest at all. The problem was kinda complex, I felt that I spent all the contest swamped in implementation, not being able to implement even 1/10 of the ideas I had. Well, it's also partially our fault for coming to troll in the finals with a 2-person team :P.

Also, did anyone manage to use the visualizer?

• +87

 » 19 months ago, # | ← Rev. 2 →   +15 We took 8th place, but results are really far from the best one, and I feel the same about "not a 4-hour contest." Scores:B — 1,070,068C — 972,941D — 1,135,201E — 739,855F — 519,975Total — 4,438,040. Initially I had and idea to start from test E, which has very small field and a lot of arms. So the idea is that in such conditions arms should work very locally and occupy as less space as possible.Pipeline of my solution is next: I put arms in some random positions, take them one by one, and completely fill each arm with tasks in some greedy order. And I visit points in the most naive way: going to the each one only by some shortest path, if it doesn't conflict with path of any of previous arms. And after visiting each assembly point I just retract the arm completely. Score function for greed ordering of tasks for current arm is $\left(estimatedLengthOfPath - \frac{score}{\alpha}\right)$. Running such solution hundreds of times with random $\alpha$ and initial order of positions I got final scores.My teammate wrote a solution with some greedy but simultaneous moving of all the arms, but it gets lower scores.
 » 19 months ago, # | ← Rev. 2 →   0 We are 32nd (team Eel) without any actual solution. We did B optimally as that was rather obvious and then got some minor really bad scores.Absolutely agree with it not being a 4-hour contest, I actually came up with some decent ideas but couldn't get them to work in time. Getting even the simplest solution one can think of to work took so long. Not happy with the problem, but I guess some people still really nailed it looking at the top scores.