AMPL GURUBI bestbound
AnsweredWhile solving mixed integer lineal models using gurobi in ampl , the bestbound reported is always 0, eaen when reaching the time limit and getting just a feasible solution, what am i doing wrong?
below the instructions used in ampl and the results display, Z is my objective function.
ampl: reset;
ampl: model ModeloMaster.mod;
ampl: option solver gurobi;
ampl: option gurobi_options 'timing 1' 'timelim=60' 'bestbound=1';
ampl: data m5n50t0.6r0.8-B-6.dat;
ampl: solve;
ampl: display Z , Z.bestbound, Z.relax, _solve_elapsed_time , solve_result ;
Gurobi 9.0.0: timing 1
timelim=100
bestbound=1
Times (seconds):
Input = 0.078125
Solve = 392.234 (summed over threads)
Output = 0.015625
Elapsed = 101
Gurobi 9.0.0: time limit with a feasible solution; objective 206
842838 simplex iterations
1843 branch-and-cut nodes
absmipgap = 206, relmipgap = 1
No basis.
No dual variables returned.
suffix bestbound OUT;
Z = 206
Z.bestbound = 0
Z.relax = 0
_solve_time = 392.344
_solve_elapsed_time = 101.344
solve_result = limit
-
Hi Hector,
Are you able to look at the full log produced by Gurobi? It seems that Gurobi does not improve the objective bound past \( 0 \) in the 60-second time limit. If you want to focus on improving the bound, a good place to start is to set the Cuts parameter to a more aggressive value and/or the MIPFocus parameter to 3. Of course, if the optimal objective value is \( 0 \), then the bound cannot be improved any further.
Thanks,
Eli
0 -
Hello,
I am having a similar problem. I am trying to find out the best bound of a simple test problem so that I can use it in my original problem. I am using the following commands on AMPL.
model 2Modefile.mod;
data 2dataset.dat;
option gurobi_options 'bestbound=1';
option solver gurobi;solve;
The output is the following.
Gurobi 9.1.1: bestbound=1
Gurobi 9.1.1: optimal solution; objective 13150
4 simplex iterationssuffix bestbound OUT;
My question is why it is showing "bestbound OUT"?
Thanks,
Pritom
0 -
I answered this question in another thread where you asked it today.
0
Please sign in to leave a comment.
Comments
3 comments