Sparkle’s documentation!¶
- 1. Sparkle user guide
- 1.1. Quick start
- 1.2. Installing Sparkle
- 1.3. Algorithm Configuration
- 1.4. Algorithm Selection
- 1.5. Executing commands
- 1.6. File structure
- 1.7. Wrappers
- 1.8. Commands
- 1.8.1. about.py
- 1.8.2. add_feature_extractor.py
- 1.8.3. add_instances.py
- 1.8.4. add_solver.py
- 1.8.5. cleanup_temporary_files.py
- 1.8.6. compute_features.py
- 1.8.7. compute_marginal_contribution.py
- 1.8.8. configure_solver.py
- 1.8.9. construct_sparkle_parallel_portfolio.py
- 1.8.10. construct_sparkle_portfolio_selector.py
- 1.8.11. generate_report.py
- 1.8.12. initialise.py
- 1.8.13. load_snapshot.py
- 1.8.14. remove_feature_extractor.py
- 1.8.15. remove_instances.py
- 1.8.16. remove_solver.py
- 1.8.17. run_ablation.py
- 1.8.18. run_configured_solver.py
- 1.8.19. run_solvers.py
- 1.8.20. run_sparkle_parallel_portfolio.py
- 1.8.21. run_sparkle_portfolio_selector.py
- 1.8.22. run_status.py
- 1.8.23. save_snapshot.py
- 1.8.24. sparkle_wait.py
- 1.8.25. system_status.py
- 1.8.26. validate_configured_vs_default.py
- 1.9. Sparkle settings
- 1.10. Required packages
- 1.11. Installation and compilation of examples
- 2. Examples
- 2.1. Use Sparkle for algorithm configuration
- 2.1.1. Initialise the Sparkle platform
- 2.1.2. Add instances
- 2.1.3. Add a configurable solver
- 2.1.4. Configure the solver
- 2.1.5. Validate the configuration
- 2.1.6. Generate a report
- 2.1.7. Run ablation to determine parameter importance based on default (from the
.pcs
file) and configured parameters - 2.1.8. Immediate ablation and validation after configuration
- 2.1.9. Run configured solver on an instance directory
- 2.2. Use Sparkle for algorithm configuration
- 2.3. Use Sparkle for a parallel algorithm portfolio with quality objective
- 2.4. Use Sparkle for a parallel algorithm portfolio with runtime objective
- 2.5. Use Sparkle for algorithm selection
- 2.5.1. Initialise the Sparkle platform
- 2.5.2. Add instances
- 2.5.3. Add solvers
- 2.5.4. Add feature extractor
- 2.5.5. Compute features
- 2.5.6. Run the solvers
- 2.5.7. Construct a portfolio selector
- 2.5.8. Generate a report
- 2.5.9. Run the portfolio selector (e.g. on a test set)
- 2.5.10. Generate a report including results on the test set
- 2.6. Use Sparkle for algorithm selection with multi-file instances
- 2.6.1. Initialise the Sparkle platform
- 2.6.2. Add instances
- 2.6.3. Add solvers
- 2.6.4. Add feature extractor
- 2.6.5. Compute features
- 2.6.6. Run the solvers
- 2.6.7. Construct a portfolio selector
- 2.6.8. [Coming soon] Generate a report
- 2.6.9. [Coming soon] Run the portfolio selector (e.g. on the test set)
- 2.6.10. Generate a report including results on the test set
- 2.1. Use Sparkle for algorithm configuration