This page displays the results of the different solvers for each benchmark for category optimisation, small integers (OPT-SMALLINT), subcategory Handmade, subsubcategory Routing
|
REMINDER
Keep in mind that the 'Best result' columns only provide the best result given by one of the solvers. This 'Best result' may be wrong in case of an UNSATISFIABLE or OPTIMUM FOUND answer (because there's no efficient way to check these answers). |
| Cell example | Meaning |
|---|---|
| Answer | Solver result |
| f=... | value of the objective function for the model reported by the solver |
| TT=... | Total Time (TT): this is the CPU time (in seconds) used by the solver until termination. This time is only meaningful for complete solvers because incomplete solvers will always run until they time out |
| T1=... | Time to get the best model (T1): this is the wall clock time (in seconds) ellapsed when the solver found its best model. This time is obtained from the 'o lines'. It is useful to compare both complete and incomplete solvers. It doesn't take into account the time to prove optimality of the solution. |
| Remember that CPU time and wall clock time are two very different notions. The CPU time represents the time during which the instructions of the solver were executed by the processor. The wall clock time represents how much time ellapsed on the clock. For a same event, the CPU time may be either smaller or greater than the wall clock time depending on the number of threads of execution and the number of processors. |
| Abbreviation | Meaning |
|---|---|
| f=... | Value of the objective function |
| TO | Time Out |
| MO | Mem. Out (out of memory) |
| Color | Meaning |
|---|---|
| text | the solver cannot handle this instance |
| text | the solver gave no answer |
| text | the solver could give an answer (SAT) |
| text | the solver gave a definitive answer (OPTIMUM FOUND or UNSAT) |
| text | the solver performed better than the other ones on that instance (complete solvers point of view) |
| text | the solver performed better than the other ones on that instance (incomplete solvers point of view) |
| text | the solver was ended by a signal or other problem |
| text | the solver gave an incomplete answer |
| text | the solver gave a wrong answer |
For better readability, you may choose to hide some solvers:
absconPseudo 1
bsolo 2006/05
glpPB 0.2
minisat+ 1.14
PBS 4.1L
PB-smodels 1.28
PB-smodels 1.31
Pueblo 1.3
Pueblo 1.4
SAT4JPSEUDO 2006.2
SAT4JPSEUDO 2006.2 Heuristics
wildcat-rnp 0.8.0
wildcat-skc 0.8.0
| absconPseudo 1 | bsolo 2006/05 | glpPB 0.2 | minisat+ 1.14 | PBS 4.1L | PB-smodels 1.28 | PB-smodels 1.31 | Pueblo 1.3 | Pueblo 1.4 | SAT4JPSEUDO 2006.2 | SAT4JPSEUDO 2006.2 Heuristics | wildcat-rnp 0.8.0 | wildcat-skc 0.8.0 | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Number of times the solver is able to give the best known answer | 5 | 15 | 15 | 15 | 9 | 5 | 5 | 13 | 15 | 10 | 10 | 0 | 0 |
| Number of times the solver is able to give the best known answer from an incomplete solver point of view (i.e. without considering optimality proof) | 5 | 15 | 15 | 15 | 9 | 7 | 6 | 15 | 15 | 14 | 14 | 15 | 15 |
| Number of times the solver is the best solver from a complete solver point of view (i.e. best known answer and best TT time) | 0 | 0 | 11 | 0 | 0 | 0 | 0 | 1 | 3 | 0 | 0 | 0 | 0 |
| Number of times the solver is the best solver from an incomplete solver point of view (i.e. best known answer and best T1 time) | 0 | 0 | 9 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 4 | 6 |