Hi all,

I'm trying to solve an MILP, but the model is not feasible. The solver is returning with the messages below. I'm trying to figure out which constraints are the culprits, but I'm unable to find out using 'getinfeas', 'getinfeascause', or 'getiis'. All three methods return empty in all cases. Is there something I'm missing? How else can I get the information about which constraints are not feasible?

thank you,

Rodrigo

solver log:

Reading Problem \xprs_71cd4c80

Problem Statistics

3653 ( 0 spare) rows

2686 ( 0 spare) structural columns

17900 ( 0 spare) non-zero elements

Global Statistics

1060 entities 0 sets 0 set members

Minimizing MILP \xprs_71cd4c80

Original problem has:

3653 rows 2686 cols 17900 elements 1060 globals

Will try to keep branch and bound tree memory usage below 22.2Gb

Starting concurrent solve with dual, primal and barrier (6 threads)

Concurrent-Solve, 1s

Dual Primal Barrier

objective sum inf

| | P 18821.000 .0000000

----- interrupted ------ | ----- interrupted ------ | ------- optimal --------

Concurrent statistics:

Dual: 1370 simplex iterations, 0.08s

Primal: 1370 simplex iterations, 0.02s

Barrier: 31 barrier and 717 simplex iterations, 0.23s

Barrier used 6 threads 4 cores, L1\L2 cache: 32K\8192K

Barrier used AVX support

Optimal solution found

Its Obj Value S Ninf Nneg Sum Inf Time

0 18820.99994 P 0 0 .000000 1

Barrier solved problem

31 barrier iterations in 1s

Final objective : 1.882099994400000e+04

Max primal violation (abs / rel) : 0.0 / 0.0

Max dual violation (abs / rel) : 9.095e-13 / 7.276e-13

Max complementarity viol. (abs / rel) : 0.0 / 0.0

All values within tolerances

Starting root cutting & heuristics

Deterministic mode with up to 8 running threads and up to 16 tasks.

Its Type BestSoln BestBound Sols Add Del Gap GInf Time

*** Search completed *** Time: 1 Nodes: 1

Problem is integer infeasible

Number of integer feasible solutions found is 0

Best bound is 18820.99994

Hi Rodrigo,

from the output it appears that you are solving an MIP, whose LP relaxation is feasible but which is integer infeasible.

Thus, it is not a single constraint that is infeasible, nor a conjunction of linear constraints, but only the linear constraints PLUS the integrality requirements together.

From the log you posted, it looks like the infeasibility was detected in cutting plane generation.

Could you specify what problems you face when trying to compute an IIS? Computing MIP-IISs can be very challenging, so is it purely a time problem or something else?

Is it possible for you to share your model with us?

Best,

Timo