Skip to content

About

An unsupervised learning method which applies a posynomial invariant form, a sum of terms each a product of powers of the metrics, to learn the shape of a trade space from evaluated designs.

Resources

Code of conduct

Contributing

Security policy

Stars

3 stars

Watchers

1 watching

Forks

Latest commit

 

History

2,416 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

tradespace

For analysts with a table of evaluated designs, to learn which metrics trade. See the docs, the paper and examples/adcs/ for a case study.

git clone https://github.com/DynamicalSystemsGroup/tradespace
cd tradespace && uv sync
# package only: uv sync --no-default-groups; uv run --no-default-groups
from tradespace import (
    HyperbolicTradeSpace,
    LogAffinePreconditioner,
    make_frontier_data,
    trade_off_report,
)

A, _ = make_frontier_data(300, 3, random_state=0)
A[:, 1] = 10 / A[:, 1]  # pretend column 1: smaller is better
M = LogAffinePreconditioner(orientation=[1, -1, 1]).fit_transform(A)
ts = HyperbolicTradeSpace(temperature=20.0).fit(M)  # real table: `select_temperature`
print(trade_off_report(ts, M).exponents.round(2))

Layout

  • src/tradespace/: the Python package (fit a frontier, read its trade-offs).
  • docs/: the user guide and API reference, built into the docs site.
  • paper/: the paper as two notebooks and what rebuilds them; site/ is the root page that links the docs and the paper.
  • examples/adcs/: the cubesat attitude-control case study the paper uses: model, data, experiments, system model.
  • derivation/: the proofs of the method's results as SymPy scripts, one per lemma.
  • tests/: the test suite; scripts/: build and check tools.
  • claims_register.toml: each claim the paper and docs make and what backs it.
  • glossary.yaml: the terms, one name per object; history.yaml: the record behind the history and acknowledgements page.
  • CHANGELOG.md: what changed, by work package.
  • work-packages/: the build's record; not needed to use the method.
  • GOVERNANCE.md: the release rules, the file that declares their settings (.github/governance.toml) and the check that runs first in CI.
  • CLAUDE.md: conventions for contributors working with Claude Code.
  • mkdocs.yml, Dockerfile, pyproject.toml, uv.lock, .python-version: build and environment files.
  • LICENSE, .github/ (CI workflows) and the dot-files .gitattributes, .gitignore, .dockerignore: license and repository settings.

About

An unsupervised learning method which applies a posynomial invariant form, a sum of terms each a product of powers of the metrics, to learn the shape of a trade space from evaluated designs.

Resources

Code of conduct

Contributing

Security policy

Stars

3 stars

Watchers

1 watching

Forks

Releases

Packages

Contributors

Languages