ICPC NERC Huawei Challenge — Cloud Scheduling Challenge
By NercNews, 4 years ago, In English

text

Hello everyone!

We are thrilled to invite you to the great event — ICPC NERC Challenge powered by Huawei 12 — 20 December 2020.

Join to Challenge

This contest is based on some real problems from Huawei Cloud team. The task is simplified, but still very challenging. You need to find a schedule for a variety of realistic problem instances, where the goal will be minimizing two different cost functions. There is a non-trivial tradeoff, decision you have to make, on how to choose what is best for a particular case. You will be provided with examples that can help you to tackle the problem. A good solution may also require diving deeper into the nature of the problem, and creating some dataset of your own. We believe there are multiple approaches one can try here, and some crossover of the algorithms will also increase your chances to win. Good luck, we hope this will be fun!

The ICPC NERC Cloud Scheduling Challenge powered by Huawei will take place online during the week of December 12-20. This round is unrated for all participants. Everybody is welcome to participate. It is an individual competition. Participants will be divided into two groups: Northern Eurasia Contestants 2020 (all subregions) and all other Participants. If you are a NERC 2020 contestant and you want to take part in ICPC NERC Challenge, link your ICPC account to the Codeforces account here: https://codeforces.me/settings/general

text

The top-scoring participants will also receive great prizes. Huawei will provide prizes to the 12 winners in both groups accordingly:

  • 1-4th place — HUAWEI P40 PRO
  • 5-8th place — HUAWEI MATEPAD PRO LTE
  • 9-12th place — HUAWEI WATCH GT 2E

In this video Andrey Tikhonov, the co-author of Cloud Scheduling Challenge will explain the basics of the task and describe some special cases and examples:

UPD: ICPC NERC Cloud Scheduling Challenge successfully finished! Thanks all for competing. Congratulations to winners with excellent results! In the video below Andrey Tikhonov will explain the problem solution:

Good luck to all participants!

  • Vote: I like it
  • +367
  • Vote: I do not like it

Comments Comments (48)
| Write comment?
»
4 years ago, # |
  Vote: I like it +9 Vote: I do not like it

Will problems be like challenge problems ,where we need to find closest approximate answer? Or the exact answers . ?

→ Reply
»
4 years ago, # |
  Vote: I like it -102 Vote: I do not like it

What is the point of participating and making prizes for 1-12 places? It's about 12 GM to participate.

I don't like it!

→ Reply
  • »
    »
    4 years ago, # ^ |
      Vote: I like it +116 Vote: I do not like it

    And how would you prefer to distribute prizes? Randomly?

    → Reply
    • »
      »
      »
      4 years ago, # ^ |
      ← Rev. 2   Vote: I like it -41 Vote: I do not like it

      No but at least giving a prize to first 50 people would have been nicer.

      → Reply
      • »
        »
        »
        »
        4 years ago, # ^ |
          Vote: I like it +62 Vote: I do not like it

        Oh, that's an awesome idea. For example, instead of awarding four smartphones to the Top-4 they can award a 25% discount coupons to the Top-16. And similarly with the other two prizes. This will surely motivate much more people to compete!

        → Reply
  • »
    »
    4 years ago, # ^ |
      Vote: I like it +30 Vote: I do not like it

    It's probably an optimization problem anyway so GM or expert, that doesn't matter.

    → Reply
»
4 years ago, # |
  Vote: I like it +19 Vote: I do not like it

Is this a marathon like contest?

→ Reply
»
4 years ago, # |
  Vote: I like it +6 Vote: I do not like it

I will be glad to participate and test my knowledge

→ Reply
»
4 years ago, # |
  Vote: I like it -196 Vote: I do not like it

Really ICPC? Giving Huaweis as prizes? Are you really that poor?

→ Reply
»
4 years ago, # |
  Vote: I like it +8 Vote: I do not like it

Awesome! Good luck to everyone participating!

→ Reply
»
4 years ago, # |
  Vote: I like it +22 Vote: I do not like it

The ICPC Huwaei Challenge is definitely one of my favorite events on Codeforces. I'm excited!

→ Reply
»
4 years ago, # |
  Vote: I like it +8 Vote: I do not like it

Good luck to everyone who is participating.

→ Reply
»
4 years ago, # |
  Vote: I like it +1 Vote: I do not like it

so is it rated? or not

→ Reply
»
4 years ago, # |
  Vote: I like it +21 Vote: I do not like it

prizes are for cf winners or just official participants of ICPC?

→ Reply
»
4 years ago, # |
  Vote: I like it 0 Vote: I do not like it

Is there any penalty for resubmitting a solutions . That is , whether should i wait for last day to submit my best solution or i can keep trying and updating.

→ Reply
  • »
    »
    4 years ago, # ^ |
      Vote: I like it +3 Vote: I do not like it

    I am sure there is no penalty for resubmitting. You can try as most as you want

    → Reply
»
4 years ago, # |
  Vote: I like it -37 Vote: I do not like it

getting bored I need Div 2 contest so bad..

→ Reply
  • »
    »
    4 years ago, # ^ |
      Vote: I like it +40 Vote: I do not like it

    Why is it boring? I find optimizations problems really interesting and fun as you have a lot of time to think and implement. It is also more realistic in real life applications as usually, you don't want a definite solution to a problem. Sometimes you only need a good approximate.

    → Reply
»
4 years ago, # |
← Rev. 2   Vote: I like it -11 Vote: I do not like it

Hey, can someone help fix the test system? I want to test deadlock so I write an infinite loop. I think it will exceed the time limit, but seems it is running endlessly. Need help thanks! UPD: never mind, it is just a little slow

→ Reply
»
4 years ago, # |
  Vote: I like it 0 Vote: I do not like it

Is this contest open to all, or only to those in certain regions or in college or university?

→ Reply
»
4 years ago, # |
  Vote: I like it +2 Vote: I do not like it

Good luck to everyone who is participating.

→ Reply
»
4 years ago, # |
  Vote: I like it +10 Vote: I do not like it

Sorry if I missed it but will our best submission be considered or the last submission??

Thanks and sorry if this was a dumb question and i missed something obvious! :D

→ Reply
  • »
    »
    4 years ago, # ^ |
      Vote: I like it +10 Vote: I do not like it

    From the problem statement:

    ... When the contest is over, then for each participant the last submission will be selected (which passed at least one test). ...

    → Reply
»
4 years ago, # |
  Vote: I like it +36 Vote: I do not like it

When is the final standings going to be published?

→ Reply
»
4 years ago, # |
  Vote: I like it +6 Vote: I do not like it

I want a matepadpro，(´;︵;`)

→ Reply
»
4 years ago, # |
← Rev. 3   Vote: I like it +10 Vote: I do not like it

What is your approach to the problem ?

Mine is keep greedily put everyone to their places if possible mainly. For each 4~5 steps try to move VMs from crowded servers to less crowded one for a few run only for special tests.

→ Reply
  • »
    »
    4 years ago, # ^ |
      Vote: I like it +18 Vote: I do not like it

    My best improvement was that it is sometimes helpful to look at the problem from the end and go to the beginning.

    → Reply
  • »
    »
    4 years ago, # ^ |
      Vote: I like it +13 Vote: I do not like it

    Greed, with order randomization and prioritizing moves from overcrowded servers. Occasionally making moves from overcrowded servers to servers with excessive cpu/ram reserves. Doing all of it until tl.

    → Reply
  • »
    »
    4 years ago, # ^ |
      Vote: I like it +3 Vote: I do not like it

    I use greedy ideas. If vms can move from an old server to a new server, I do it. If count of moves in step less than some constrain, I move vms from servers that haven't all necessary vms, to servers that already have all necessary vms. This move ensure that any vm would move no more than 2 times. Practice shows that no situation when such move not exist in these testcases. Also I use some randomized algorithms and optimizations.

    → Reply
  • »
    »
    4 years ago, # ^ |
      Vote: I like it +3 Vote: I do not like it

    First I implement a naive greedy algorithm which simply moves the vms without turning servers. It may causes some deadlock, so during some steps I will check every server and use some of them with large space and small moves as turning points. Finally I set up 10+ sets of parameters (like the frequency of using turning servers, how to choose turning servers, etc), run them seperately and use their minimum as my answer.

    It's also important to use randomization and time trick (continuous running until TL).

    → Reply
»
4 years ago, # |
  Vote: I like it 0 Vote: I do not like it

I have struggled a lot for winning a HUAWEI MATEPAD PRO LTE. The standing shows that many participants are stronger than me >_<. btw, will others' submissions be made public?

→ Reply
»
4 years ago, # |
  Vote: I like it +73 Vote: I do not like it

A small suggestion for improvement such type of problems.

Since your solution is tested only on 30 tests of 50 during the round, there is no guarantee that this solution will work correctly on the other 20 tests. So you have to create more "careful" solutions. In particular I "asked" my program to run only 4 seconds (while 5 or even 5.5 seems to be OK too) and have removed some optimizations. But I was not sure that my solution works properly till end of the system test, because code is very large and complex. Also the time measurement is a bit unstable, so you have to be very very careful.

From the other hand, I enjoy the CodeChef challenging problems system:

Quote

Once you may test you solution fully, you may focus on new optimizations/ideas. May be it will be good to implement something like this here.

BTW, thanks for the great competition!

→ Reply