integer infeasible, No basis.

[Saber Jalilzadeh-Galaeh]

Hi All,

I have a MIP model and trying to solve it using CPLEX.

I have the following error.

Anybody knows why this happens and who I could get ride from this bug?

by the way, model and data files are attached. 

Thanks in advance

Saber

650 variables, 665 constraints, 1 objectives

presolve results:

variables omitted: 280

constraints omitted: 455

ranges relaxed: 0

bounds improved: 615

nba cycles: 1

bound row scans: 260

row-scan updates: 200

Presolve eliminates 455 constraints and 480 variables.

Adjusted problem:

170 variables:

20 binary variables

150 linear variables

210 constraints, all linear; 800 nonzeros

1 linear objective; 82 nonzeros.

CPLEX 12.3.0.0: mipgap=0

mipdisplay=4

timing=1

integrality=0

MIP Presolve eliminated 35 rows and 20 columns.

Reduced MIP has 165 rows, 140 columns, and 700 nonzeros.

Reduced MIP has 10 binaries, 0 generals, 0 SOSs, and 0 indicators.

Probing fixed 4 vars, tightened 30 bounds.

Probing time =    0.00 sec.

Times (seconds):

Input =  0.002

Solve =  0.01

Output = 0.004

CPLEX 12.3.0.0: integer infeasible.

0 MIP simplex iterations

0 branch-and-bound nodes

Tried aggregator 2 times

No basis.

[Paul]

It's not a bug; your model is infeasible (as a MILP; the LP relaxation is feasible, with fractional values for x).

To diagnose the infeasibility, add the following line to the model before solving it:

 option cplex_options "iisfind=1";

When you solve the model, you'll see CPLEX's progress looking for an IIS.  When it's done looking, do the following:

display constraint_1.iis;
display constraint_2.iis;
...

Look at the constraints that are flagged as contributing to infeasibility (in combination with the integrality restriction on x) and see if you can sort out what you did wrong.

Cheers,
Paul

[Robert Fourer]

To get a convenient list of only the components of an irreducible infeasible
subset (IIS), you can use the following two commands:

display {i in 1.._ncons: _con[i].iis <> 'non'} (_conname[i],_con[i].iis);
display {j in 1.._nvars: _var[j].iis <> 'non'} (_varname[j],_var[j].iis);

The presence of a variable in the listing implies that the constraint
imposed by a bound on the variable is part of the IIS.

Bob Fourer
4...@ampl.com

[Saber Jalilzadeh-Galaeh]

Dear Paul, Dear Bob,

Thanks a lot for your helps-

Saber

---

Von: Robert Fourer <4...@ampl.com>
An: am...@googlegroups.com
Gesendet: 1:32 Samstag, 13.August 2011
Betreff: RE: [AMPL 4918] Re: integer infeasible, No basis.

--
You received this message because you are subscribed to the Google Groups "AMPL Modeling Language" group.
To post to this group, send email to am...@googlegroups.com.
To unsubscribe from this group, send email to ampl+unsub...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/ampl?hl=en.