Re: Regression between 4.61 and 4.65 for problems which can be preprocessed to almost nothing

[Heinrich Schuchardt]

On 13.07.20 17:20, Domingo Alvarez Duarte wrote:
> Hello Peter !
>
> Till glpk 4.64 this error does not happen, I'm looking through the code
> and think that the real fix is to prevent to solve the problem if the
> presolver eliminated all columns or rows.
>
> Cheers !
>

The presolver may find an integer optimal solution. Then the solver
stops with:
INTEGER OPTIMAL SOLUTION FOUND BY MIP PREPROCESSOR

Currently this is only reached if the transformed equation system has
zero columns and zero rows (src/draft/glpapi09.c, line 328). In the
example the presolved equation system has 0 columns and 1 row.

Why is this row is not active and therefore not eliminated in
npp_process_prob() (src/npp/npp5.c:651)?

Best regards

Heinrich