Good Afternoon I am experimenting with a fortran-glpk interface for a problem that requires solving several tens (or hundreds) of thousands of slightly modified LP problems. The fortran-glpk interface has substantially reduced my computation time as contrasted to running my code in R via the lpsolveAPI or glpkAPI package. For well behaved problems, the fortran-glpk interface solves the set of problems in 1/3 to 1/6 the time required to solve the problems using R's lpsolveAPI. With this application, R's glpkAPI is not time competitive with lpSolve.
However, in my application there will be times when my problem will be primal unbounded. I have a small demonstration example where I know the problem is unbounded and is identified as such by both lpSolve and MS excel's solver. However, when I attempt to solve this problem using the fortran-glpk interface, the glp_simplex call returns 0. How do I identify an unbounded problem when calling glpk via the glp_simplex function call. I have read the GNU Lineal Programming Kit Reference Manual and have not found (or not understood) how to identify an unbounded problem. Thank you for your assistance Joe Atwood Montana State University Bozeman, MT 59715 406-994-5614
_______________________________________________ Help-glpk mailing list [email protected] https://lists.gnu.org/mailman/listinfo/help-glpk
