Call for Participation
ARCH-COMP 2027 — the 11th International Competition on Verifying Continuous and Hybrid Systems — invites tool developers and benchmark proposers across all seven categories. The competition is friendly: within each category, participants jointly agree on benchmarks and metrics, then run their tools on the shared instances.
The Process
Join a category group
Pick the category (or categories) that treat problems similar to yours and join that group. Each category is coordinated by a lead who can point you to the current benchmarks and discussion. New participants are always welcome.
Agree on benchmarks & submit preliminary results
Each group settles its benchmark set — everybody can propose, and all should approve. Benchmarks already in the ARCH repository are preferred; problems are refined in group discussion. Models and data are collected under goranf/ARCH-COMP/models.
Submit your repeatability package
Provide install and run scripts (containerized with Docker is preferred) so your results can be independently reproduced. Submit through the evaluation platform.
Prepare the results report
Each category assembles its results report. The evaluation chair confirms that the repeatability packages reproduce the reported results and signals approval.
Present & publish
Results are presented at the ARCH 2027 workshop and published as a category report in the EPiC Series in Computing proceedings (DBLP-indexed).
Who can take part
Tool Developers
Enter your verification tool in one or more categories and compare against the state of the art on agreed benchmarks.
- Pick your categories and join the discussion
- Provide install & run scripts (Docker preferred)
- Submit a repeatability package
Benchmark Proposers
Contribute challenging, real-world benchmarks that push the tools in a category forward.
- Coordinate with the category leads
- Provide models, specifications, and instances
- Document the expected verification results
New Categories
ARCH-COMP evolves with the community. Proposals for new problem classes are welcome.
- Sketch the problem class and metrics
- Line up an initial set of participants
- Contact the general chairs