Sebastian Erik Løken Rindvoll

Gurobi-versary
Conversationalist
Curious
  • Total activity 30
  • Last activity
  • Member since
  • Following 0 users
  • Followed by 0 users
  • Votes 1
  • Subscriptions 6

Activity overview

Latest activity by Sebastian Erik Løken Rindvoll

  • Sebastian Erik Løken Rindvoll created a post,

    How Gurobi's concurrent MIP solver works

    Answered
    1. Community
    2. Algorithms

    Hi!  I am using your concurrent MIP solver in PyCharm to solve my fleet size and mix problem for the installation process of floating offshore wind. When I increase the number of turbines to be ins...

    • Sebastian Erik Løken Rindvoll
    • 2 followers
    • 1 comment
    • 0 votes

  • Sebastian Erik Løken Rindvoll commented,

    1. Community
    2. Other
    3. Same incumbent solution all the time

    Tobias Achterberg But can you tell anything from the log provided below? Can I see already from the presolve process that I have chosen the wrong upper or lower bounds?   /Users/sebastianeriklokenr...

    • View comment
    • Sebastian Erik Løken Rindvoll
    • 0 votes

  • Sebastian Erik Løken Rindvoll commented,

    1. Community
    2. Other
    3. Same incumbent solution all the time

    Thank you very much Tobias Achterberg. But how do I provide a MIP start in PULP in PyCharm. Right now I am using this:   prob.solve(GUROBI(MIPGap=0.01, Presolve=2, Cuts=2, Symmetry=2))

    • View comment
    • Sebastian Erik Løken Rindvoll
    • 0 votes

  • Sebastian Erik Løken Rindvoll commented,

    1. Community
    2. Other
    3. Same incumbent solution all the time

    I have not completely understood if I can provide the model with a solution from the start to make sure that the lower bound does not start as low as 8448. Or is it possible to make the model solve...

    • View comment
    • Sebastian Erik Løken Rindvoll
    • Edited
    • 0 votes

  • Sebastian Erik Løken Rindvoll commented,

    1. Community
    2. Other
    3. Same incumbent solution all the time

    With different seeds, I have noticed that the solution appears at different times. The memory on the windows computer have never been an issue, but the CPU varies from 6% to 70% during the whole se...

    • View comment
    • Sebastian Erik Løken Rindvoll
    • 0 votes

  • Sebastian Erik Løken Rindvoll commented,

    1. Community
    2. Other
    3. Same incumbent solution all the time

    I have tried with different seeds, and it seems like that is the case, but I agree with you that it should not be like that. I have seen that I only use about 20% of the CPU on the windows computer...

    • View comment
    • Sebastian Erik Løken Rindvoll
    • 0 votes

  • Sebastian Erik Løken Rindvoll commented,

    1. Community
    2. Other
    3. Same incumbent solution all the time

    The Windows computer has more computational power than the mac, but the Mac seems to find solutions and explore nodes much faster than the Windows computer.

    • View comment
    • Sebastian Erik Løken Rindvoll
    • 0 votes

  • Sebastian Erik Løken Rindvoll commented,

    1. Community
    2. Other
    3. Same incumbent solution all the time

    This is my Windows Log file:   C:\Users\serindvo\PycharmProjects\untitled\venv\Scripts\python.exe W:/PycharmProjects/untitled/venv/GurobiModel.pyUsing license file C:\Users\serindvo\gurobi.licAcade...

    • View comment
    • Sebastian Erik Løken Rindvoll
    • 0 votes

  • Sebastian Erik Løken Rindvoll commented,

    1. Community
    2. Other
    3. Process finished with exit code 137 (interrupted by signal 9: SIGKILL)

    Thank you! :)  But it is difficult to test for my model since when I increase one of the parameters in my model from 3 to 4 it goes from 30 seconds to over 20 000 seconds. And if I have understood ...

    • View comment
    • Sebastian Erik Løken Rindvoll
    • 0 votes

  • Sebastian Erik Løken Rindvoll commented,

    1. Community
    2. Other
    3. Same incumbent solution all the time

    And how can it be that the windows computer searches through nodes faster than the mac computer the first thousand seconds? After a thousand seconds the Mac computer searches through nodes much fas...

    • View comment
    • Sebastian Erik Løken Rindvoll
    • 0 votes