Should I be interested in this integer problem? #558
Replies: 3 comments 1 reply
-
|
Fscip from Scip suite 8.0.4 found -10507 in a few seconds with 11 threads. Or is this already solved? $ cat mas76mod.sol |
Beta Was this translation helpful? Give feedback.
-
|
Impressive. Did it show any signs of finishing soon? |
Beta Was this translation helpful? Give feedback.
-
|
The job died after ~2 hours, I think it was out of memory. I have 40 GB mem. I will try again with less threads and such .. |
Beta Was this translation helpful? Give feedback.
-
|
Using that -10507 solution as a mipstart, Neos Gurobi found -10508 263136355 2464513 -10512.000 59 7 -10507.000 -10512.000 0.05% 3.0 4240s |
Beta Was this translation helpful? Give feedback.
-
This is not a very serious discussion, but it would be very interesting if someone could solve the problem quickly.
I look at Cbc's preprocessing and try and see what I can do to improve it. I have been looking at various ideas which include altering the model by scaling and some simple modifications. Sometimes I use a miplib example and fix a few variables to make it easier. Some months ago I played around with the miplib problem mas76 and ended up with a problem of 60 0-1 variables and 7 rows (and 100% dense!). This was meant to be easier to solve than the original!!
I attach the problem
mas76mod.mps.txt
However this problem does not seem easy to solve. The continuous relaxation is -10512. I have given it to a good commercial solver and it has got a solution of -10505, but I am not sure it will finish very soon. It has been running for two days and done 1.2e9 nodes! More worryingly, it is slowing down.
Anyway - do people think I should offer it to the miplib test set or should I just think that anyone can invent a difficult to solve problem?
Beta Was this translation helpful? Give feedback.
All reactions