CPLEX return error solution

[riccardo zanella]

Good evening!
I have implemented a model but when i solve them:

ampl: model '/home/riccardo/amplide-demo/example/Second_problem.mod';
ampl: option solver cplex;
ampl: option cplex_option 'primalopt';
ampl: solve;
CPLEX 12.6.1.0: unbounded problem.
19 dual simplex iterations (19 in phase I)
variable.unbdd returned
2 extra simplex iterations for ray

suffix unbdd OUT;
ampl: display u_0;
u_0 = 0

ampl: display u_1;
u_1 = -1.14573

ampl: display u_7;
u_7 = 1.75

ampl: display a_0;
a_0 = -0.25

this solution isn't correct in fact (for exemple) daesn't respect my first constraint:

subject to Constraints1: - u_0 - u_1 - u_7 + a_0 >= 0;

-0 +1.14573 -1.75 -0.25 <0   <=>  -0.854272 < 0

why?? where i wrong?

[riccardo zanella]

possible that the problem is that I use dual simplex?

ampl: model '/home/riccardo/amplide-demo/example/Second_problem.mod';
ampl: option cplex_option 'primal';

ampl: solve;
CPLEX 12.6.1.0: unbounded problem.
19 dual simplex iterations (19 in phase I)
variable.unbdd returned
2 extra simplex iterations for ray

suffix unbdd OUT;

if the answer is yes why?

[riccardo...@gmail.com]

Good evening!
I have a problem that I can't resolve!

I implemented a simple model the output is:

ampl: model '/home/riccardo/amplide-demo/example/Second_problem.mod';

ampl: option solver cplex;
ampl: option cplex_option 'primalopt';

ampl: solve;
CPLEX 12.6.1.0: unbounded problem.
19 dual simplex iterations (19 in phase I)
variable.unbdd returned
2 extra simplex iterations for ray

suffix unbdd OUT;

ampl: display u_0;
u_0 = 0

ampl: display u_1;
u_1 = -1.14573

ampl: display u_7;
u_7 = 1.75

ampl: display a_0;
a_0 = -0.25

This isn't a correct solution in fact  (for example)  daesn't respect my first constraint:

subject to Constraints1: - u_0 - u_1 - u_7 + a_0 >= 0;

-0+

1.14573-1.75-0.25 < 0 => -0.854272 < 0

I don't understand why... Where I am wrong??
I


[Robert Fourer]

CPLEX reports an "unbounded problem" so you know that an optimal solution does not exist, and the solution that CPLEX returned is not an optimal one. But since CPLEX used the dual simplex method, it doesn't even have a feasible solution to return. To see whether there is a feasible solution you should try running the primal simplex, which you tried to do, but your specification should have been

option cplex_options 'primalopt';

(that is, with "options" rather than "option").

Bob Fourer
am...@googlegroups.com

=======