[Coin-lpsolver] Interpretation of columnLower() and columnUpper() output in drive r2

[Brady Hunsaker]

Vansant, Robert wrote:
> New user question...
> 
> If I run the COIN/Clp/Samples/driver2 example (Nov16 build) on a simple
> example:
> 
> NAME MYTEST
> ROWS
>   N 1
>   E 2
> COLUMNS
>     X         2            1.0000000
>     X         1            2.0000000
>     Y         2            1.0000000
>     Y         1            1.0000000
> RHS
>     RHS       2            1.0000000
> ENDATA
> 
> I get the, at the end, the output:
> 
> Clp0000I Optimal - objective value 1
> mytest.mps Objective 1 took 0 iterations and 0 seconds
> Model MYTEST has 1 rows and 2 columns
>                        Primal          Dual         Lower         Upper
> (Cost)
>      0        2      1.000000      1.000000      1.000000      1.000000
> --------------------------------------
>                        Primal          Dual         Lower         Upper
> Cost
>      0        X      0.000000       1.00000       0.00000  1.79769e+308
> 2.00000
>      1        Y       1.00000       0.00000       0.00000  1.79769e+308
> 1.00000
> --------------------------------------
> 
> My interpretation of Lower and Upper bounds (columnLower() and columnUpper()
> calls) is that they indicate/return the bounds on the objective coefficients
> that would cause the variable to enter or leave the basis, but that doesn't
> seem consistent with the column outputs above.  What are these column ranges
> indicating?
> 
> Many Thanks,
> Bob Van Sant

You are expecting sensitivity bounds--perhaps because Excel Solver has
output like that.   That's not what CLP is giving you.  It's just
reminding you what the bounds on the variables are in the model.  So
both X and Y are nonnegative with no upper bound (that's infinity).

Brady

-- 
Brady Hunsaker
Assistant Professor
Industrial Engineering
University of Pittsburgh
http://www.engr.pitt.edu/hunsaker/