Jan Krause
- 合計アクティビティ 13
- 前回のアクティビティ
- メンバー登録日
- フォロー 0ユーザー
- フォロワー 0ユーザー
- 投票 0
- サブスクリプション 1
コメント
Jan Krauseによる最近のアクティビティ-
Thanks
-
It seems that - the presolved model (having Presolve=0) has R0+all Mccormick envelopes (explicitly) + all product variable equations + trivial bounds 0 and 1 - my manual McCormick relaxation has R0...
-
Yes, that is correct. Thanks for your help!
-
Here is the .lp file \ LP format - for model browsing. Use MPS format to capture full model detail.Maximize 1.29 x[1] + 3.91 x[2] - 1.24 x[3] - 0.68 x[4] - 2.74 x[5] - 2.09 x[6] + 4.72 x[7] - 1....
-
It turned out, that even only R0: x[1] + x[2] + x[3] <= 0.48 yields a difference in the values.
-
In this case the values are the same.Jaromił Najman I have tested the same problem (with all constraints) in SCIP. There I explicitly disabled propagation and then the values are the same. It woul...
-
-
Yes. When I solve with Presolve=0 and NodeLimit=0, such that the optimization terminates with the root relaxation without cuts (is that right?), then I get an objective value 4.381421504525e+01, th...
-
Hi Jaromil, thanks for your quick response. It would be interesting for me which of these trivial bound reductions are applied in my specific case. In my instance, I have 15 x-variables, 15 y-varia...
-
Hello Jaromił Najman, I found that the root relaxation excluding Presolve is apparently not the same as the relaxation that is obtained by just adding the McCormick envelopes. I assume that Gurobi ...