Empirical performance models
can be used to predict the performance of algorithms on previously
unseen input, including previously unseen
problem instances, previously untested parameter settings, or both.
- Frank Hutter, Holger Hoos,
Key Algorithm Parameters and
Instance Features using Forward Selection
Here is a zip
file with the algorithm
underlying our AIJ submission.
It includes the runtimes and
for our 35 benchmark distributions, as well as the runtime matrices for
CPLEX and Spear.
Here is a zip
file with the Matlab code
we used in our
experiments. Note that this zip
file does not include the source
networks, approximate Gaussian processes, or SPORE-FoBa due to
The main file for experiments for the feature space is
epm_experiments.m; the main file for experiments in the configuration
space & the joint space is epm_matrix_experiments.m; see the
README.TXT for their usage.
Here is our feature
computation code for SAT, feature
computation code for MIP,
code for TSP. If you use
these features, please cite the AIJ
We also have a modified version of the SAT feature computation code, which only computes some quite cheap features, which we used to generate features for the Configurable SAT Solver Competition (CSSC).
For TSP, you might also want to obtain the feature
computation code of Smith-Miles
and van Hemert.
Our submission to LION-7 extends the methodology of EPMs to perform
forward selection for identifying key parameters and features. Here is
code; the data remains unchanged
Please send any questions,
concerns or comments to Frank