Robotics · sim sweep design & coverage

domain-rand-sweep

A sweep-design and coverage tool that compares exhaustive grid configurations with a smaller Latin-hypercube plan before simulator compute is committed.

Public · synthetic demo
The problem

A large sweep can waste compute without proving useful coverage.

Sampling plans should be inspected for coverage and duplicates before they become simulator jobs.

The money shot

Grid 1215 configurations vs Latin hypercube 64

Measured sweep plans
planconfigurations
grid1215
Latin hypercube64
0exact duplicate groups
0near-duplicate groups
20240101seed

The configuration count falls to roughly one nineteenth of the grid while duplicate groups remain absent.

Honest limitations

This tool does not run a simulator, train or evaluate a policy, or claim that sweep coverage transfers to the real world. Broad coverage is not evidence of real-world robustness.

This page reports the supplied sample output and its stated limits. Inspect the code and evidence ↗