Infeasible Stochastic Model

GAMS
1

For some reason, when I run the model the log returns:

Reading data...
Starting Cplex...
Unable to load names.
Version identifier: 12.10.0.0 | 2019-11-26 | 843d4de2ae
CPXPARAM_Advance                                 0
CPXPARAM_Threads                                 1
CPXPARAM_Parallel                                1
CPXPARAM_Simplex_Limits_Iterations               2000000000
CPXPARAM_TimeLimit                               1000
CPXPARAM_Tune_TimeLimit                          200
Row 'c1339' infeasible, all entries at implied bounds.
Presolve time = 0.00 sec. (0.56 ticks)
LP status(3): infeasible
Cplex Time: 0.00sec (det. 0.57 ticks)
Rerunning with presolve turned off and solving with LPMETHOD=0
Version identifier: 12.10.0.0 | 2019-11-26 | 843d4de2ae
CPXPARAM_Advance                                 2
CPXPARAM_Preprocessing_Presolve                  0
CPXPARAM_LPMethod                                1
CPXPARAM_Threads                                 1
CPXPARAM_Parallel                                1
CPXPARAM_Simplex_Limits_Iterations               2000000000
CPXPARAM_TimeLimit                               1000
CPXPARAM_Tune_TimeLimit                          200

Iteration log . . .
Iteration:     1    Scaled infeas =         37102.517619
Iteration:    60    Scaled infeas =         37084.090129
Switched to devex.
Iteration:   527    Scaled infeas =         33928.201929
LP status(3): infeasible
Cplex Time: 0.02sec (det. 8.77 ticks)

Model has been proven infeasible.

Could anyone help me?

(I’m brazilian, sorry about my english)
programa_teste2-emp.gms (4.69 KB)

2

Without the Excel file I could not reproduce your findings. First thing I would do to see if the deterministic model is feasible. So before you do some fancy stochastic stuff, solve the deterministic model as a regular LP and see if this is feasible or not. It’s probably much easier debugging this instead of the stochastic model.

defineXs0(i)..          Xs(i, '1') =e= 0;
c.lo(t) =  clo;
c.up(t) =  cup;
model ALM_test /all/;

solve ALM_test min zmaster using lp;
$exit
* EMP Annotations
File emp / '%emp.info%' /;

-Michael

3

Sorry, I forgot the Excel file. Now is avaliable, and (at least what I know, because the model comes from a paper) the model is feasible. But something is wrong in the implementation… Also, I want to know where I can see the “Row ‘c1339’”!!!
Liability.xlsx (9.89 KB)

4

As I suspected, the deterministic problem is already infeasible:

...
CPXPARAM_Tune_TimeLimit                          200
Infeasibility row 'cashbalance(1)':  0  = -11.4128.
Presolve time = 0.00 sec. (0.01 ticks)
LP status(3): infeasible
Cplex Time: 0.00sec (det. 0.01 ticks)
...

There are plenty of ways to deal with infeasible models. Since this seems to be a model from the literature, there is probably a chance that the authors reported the solution. Just set the solution before the solve (using variable.l(…) = …; syntax, turn on full equation listing (modelname.limrow=1e9;) and solve. In the equation listing search for INFES which give you the equations that are infeasible in your implementation.

-Michael

5

Thanksss!!!