| Name | noswot |
| Download | noswot.mps.gz |
| Solution | noswot.sol.gz |
| Orginator | |
| Formulator | Linus E. Schrage |
| Donator | John W. Gregory |
| Rows | 182 |
| Cols | 128 |
| Non-zeros | 735 |
| Integers | 25 |
| Binaries | 75 |
| Continuous | |
| |Min| | 2.50000000e-01 |
| |Max| | 2.10000000e+01 |
| Integer Objective | -4.10000000e+01 |
| LP Objective | -4.30000000e+01 |
| Root LP Basis | noswot.bas.gz |
| Set partitioning | |
| Set packing | |
| Set covering | |
| Cardinality | |
| Equality Knapsacks | |
| Bin packing | |
| Invariant Knapsack | |
| Knapsacks | |
| Integer Knapsack | 32 |
| Upper bounds | |
| Lower bounds | 20 |
| Mixed 0/1 | |
| General Cons. | 130 |
| References |
With a little study of the model, the problem becomes easy.
The summary description of what can be done can is given on pages 25 and 26 of the paper
IP optimum found by CPlex 9.0 in 14.3 CPU hours.
noswot solves in well
under an hour, but that was with a tuned setting. Default settings
took more than an hour. In fact, it would be easy to conclude that
noswot wasn't going to solve since the bound didn't move for almost an
hour. However, there is a hint in the fact that the LP values
usually differ from the bound.