MaxSAT Evaluation 2017
Affiliated with SAT 2017   ·   August 28 - September 1   ·   Melbourne, Australia

Tracks

Solvers can be submitted for evaluation in the following tracks, characterized by the types of benchmarks instances used within each track to evaluate the solvers. Please note that MSE 2017 does not include a track for randomly generated MaxSAT instances.

Main Tracks

  • Unweighted: Combines the industrial and crafted unweighted and unweighted partial MaxSAT categories from previous MaxSAT evaluations. Purely randomly generated instances are not included.
  • Weighted: Combines the industrial and crafted weighted and weighted partial MaxSAT categories from previous MaxSAT evaluations. All benchmarks will be truly weighted, i.e., contain soft clauses with different weights. Purely randomly generated instances are not included.

Special Tracks

  • Incomplete Special Tracks: Two special tracks, unweighted and weighted (following the categorization of the main tracks), for incomplete solvers are organized. Every incomplete solver participating in one or both of the tracks will be evaluated using two distinct per-instance time limits: 60 seconds and 300 seconds. The ranking criterion for these tracks is described under rules.
  • No-restrictions Track: This special track allows solvers not adhering to the open source requirements of MSE 2017 to take part in the evaluation. Portfolios applying instance-specific algorithm selection technique to select an third-party MaxSAT solver to run on a per-instance basis are also primarily encouraged to participate in this track. Similarly as in the other tracks, solvers should make use of only one processor core. Benchmarks for the no-restrictions track consist of a mixture of the benchmarks in the unweighted and weighted main tracks.