1266890766

Source: Dattorro.lp
File: upload/23.02.2010_1266890766_Jon_Dattorro/Dattorro.lp
Name: Jon_Dattorro
Company: convexoptimization.com
eMail: dattorro@stanford.edu
Street: PO Box 12
City: Stanford, CA 94302
Country: UNITED STATES (US)
Message:   minimize  1\'x
subject to  A x = b
            x in {0,1}^n

This integer program is submitted to MIPLIB 2010, in LP format, at the request of Robert Bixby in Feb.2010.

It expresses a practical problem having real-world application; it is not a randomly generated problem.

Matrix A is sparse having integer entries in {-1,0,1,2} and dimensions 11,077 x 262,144.  

Only 0.05% of the entries in A are nonzero.  

Vector b is sparse in {0,1}^n.

A good commercial presolver can eliminate at least 50,000 columns and about 1,000 rows. 

A desired optimal solution  x  is binary with minimum cardinality: 256.  

The given objective is simply a sum of all variables, 1\'x ;  but it is actually arbitrary.

Any binary solution to  Ax=b  having cardinality 256 is desired.

Jon Dattorro

delivered on: 23.02.2010
Fileupload successfully confirmed!