NEOS Server Home

NEOS Interfaces to feaspump

WWW Form & Sample Submissions
E-mail
XML-RPC

feaspump





The NEOS Server offers FEASPUMP for the solution of mixed-integer linear feasibility problems in AMPL format. See these notes on formats and conversion. FEASPUMP stops after having found one feasible point. As an alternative you can use SCIP with limits/solution = 1

The feasibility pump (FP) algorithm was originally developed by Fischetti, Glover, and Lodi for finding feasible points of binary LPs. It was then generalized to mixed-integer LPs by Bertacco, Fischetti, and Lodi ("BFL"). Achterberg and Berthold ("AB"; ZIB Berlin) developed this algorithm further. FP uses the CPLEX library. Source code and literature are available here.

This implementation permits to call the original BFL version of FP or the AB version. For a feasibility benchmark, see here.

This solver was implemented by Hans Mittelmann and executes in 64 bit under Linux at


Using the NEOS Server for FEASPUMP

The user must submit a mixed-integer linear programming problem in AMPL format. Examples of models in AMPL format can be found in the MIPLIB 2003 library.

Only a model and a datafile may be submitted no command file. The reason is that an MPS file is produced from the AMPL model and a file is put out that shows the correspondence of the variable names in the AMPL script and the MPS file respectively.
Note: Do not use an "end" at the end of the model file!
Note: Problems must be formulated as minimization problems!

The administrator of FEASPUMP reserves the right to keep and use problem instances submitted through NEOS for testing and debugging purposes.


Enter the complete path to the AMPL model file (linear, minimization, no commands, no "end")
AMPL model:


Enter the complete path to the AMPL data file (optional, no commands)
AMPL data:


Choose the FP version (ab, bfl)
FP version:



Comments:


e-mail address:

Please do not click the 'Submit to NEOS' button more than once.


NEOS Server Home
Submit comments and questions


DOE disclaimer
DOE Web privacy policy