Please, try EDU on Codeforces! New educational section with videos, subtitles, texts, and problems. ×

Recent actions

And will you pay the import tax and duties? ICPCNews1 Last time I got a prize from a Huawei contest, I was billed 189 USD in import tax and administration fees from DHL after receiving the package. There's 25% import tax in Norway, and DHL sends the bill to the receiver if the sender doesn't take care of it. I might decline the prize rather than paying the bill.

Created or updated the text

how soon winners will get their prizes?? ICPCNews1

Can you please keep a version of ICPC open for practice .

Thanks a lot!

I published post with some ideas link

I just used an existing implementation and added some extra bookkeeping data to be able to recompute the regularization quickly. The $$$\gamma$$$ is the "resolution" and is just a parameter for the objective function (for modularity, $$$\gamma = \frac1{2|E|}$$$).

For the local search, let $$$S$$$ be the current best score and $$$S'$$$ be the score after a proposed change. If $$$S > S'$$$, my algorithm would accept the change with probability $$$\min(0.02, \exp(10(S'-S))$$$. The proposed change was the best change among all possible moves for 55 randomly sampled nodes. There's no reason for choosing any of these parameters, it just worked well enough in practice.

What does your function for MergeNodesSubset for Leiden look like? Did you use the γ in the pseudo-code and what is that, is it a function?

Also what was your initial temperature for SA and how did you decrease the temperature, I didn't get great scores with it compared to hill climbing and if I did it took a long time. I used temperature = 1 and decreased it linearly by .1 or .01 (and my formula for the probability looked like this: exp(delta*10000/temperature)).

For a simpler solution, I found that greedily creating singleton clusters to maximize regularization on the partition given by the original Leiden algorithm works well also.

My best solutions came from applying the Leiden algorithm adapted to fit the objective function, coupled with some local search/simulated annealing to get a few dozen extra points.

The contest is over now, so, can someone from top share main ideas of algorithm? Thanks in advance

Can i solve problems after ICPC Challenge 2020 and ICPC Challenge 2020: Marathon?

Created or updated the text