[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Bug-glpk] [Fwd: BUG report]
From: |
Andrew Makhorin |
Subject: |
Re: [Bug-glpk] [Fwd: BUG report] |
Date: |
Mon, 27 May 2013 22:22:51 +0400 |
> I think that I found a BUG.
>
> LP problem
>
>
> Maximize
> obj: + 3 Lanac_1 + 5 Lanac_2 + 4 Lanac_3 + 4 Lanac_4 + 3 Lanac_5
>
> Subject To
> C6: + Lanac_5 + Lanac_3 + Lanac_1 <= 1000
> C7: + Lanac_3 + Lanac_2 + Lanac_1 <= 560.5
> C13: + Lanac_3 + Lanac_2 <= 2500
> C12: + Lanac_4 + Lanac_2 <= 855.2
> C15: + Lanac_4 + Lanac_2 <= 800
> C16: + Lanac_4 + Lanac_2 <= 1000
> C14: + Lanac_5 + Lanac_3 <= 700
> C11: + Lanac_5 + Lanac_4 <= 254.8
>
> Bounds
> 200 <= Lanac_1 <= 560.5
> 200 <= Lanac_2 <= 560.5
> 200 <= Lanac_3 <= 560.5
> 200 <= Lanac_4 <= 254.8
> 200 <= Lanac_5 <= 254.8
>
> End
>
> For all my variables I'm getting 200 as a result which is not
> correct if we look at constraint C7 or C11:
>
> C7: + Lanac_3 + Lanac_2 + Lanac_1 <= 560.5
>
> 600 is not less than 560.5
Your instance is primal infeasible, i.e. it has no primal feasible
solution; you must see a diagnostic message about that. In this case the
glpk simplex solver returns a basic solution that minimizes the sum of
infeasibilities, and since that solution is primal infeasible, it
violates some constraints.