NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
PyPI
PyPI
Design of Experiments for Python
Latest stable
1.6.0
Last release 1 months ago
21 Oct 2026
Ships unpredictably
gaps range from 2 weeks to 10.6 years
Nearly every release is documented
notes for 41 of 45 stable releases
Nothing withdrawn
no release was ever pulled
13 years old
46 releases · first in 2013
Drag across the ruler to compare, or tap a release to open it.
Check dependencies monthly instead of weekly by @saudzahirr in https://github.com/pydoe/pydoe/pull/166
Full Changelog: https://github.com/pydoe/pydoe/compare/v1.5.0...v1.6.0
Bump ruff from 0.16.0 to 0.16.1 by @dependabot[bot] in https://github.com/pydoe/pydoe/pull/150
fracfact_opt return saturated fractional factorial designs by @saudzahirr in https://github.com/pydoe/pydoe/pull/156fracfact_by_res designs that actually reach the requested resolution by @saudzahirr in https://github.com/pydoe/pydoe/pull/157Full Changelog: https://github.com/pydoe/pydoe/compare/v1.4.0...v1.5.0
Your coding agent can read these notes before it upgrades. Set up the MCP server →
Update docs by @saudzahirr in https://github.com/pydoe/pydoe/pull/130
Full Changelog: https://github.com/pydoe/pydoe/compare/v1.3.0...v1.4.0
Implementation v1.3.0 by @saudzahirr in https://github.com/pydoe/pydoe/pull/129
Full Changelog: https://github.com/pydoe/pydoe/compare/v1.2.0...v1.3.0
Maximum projection design ( maxpro_design ) — optimizes Latin hypercube projections through coordinate-exchange search — @saudzahirr
Nearly orthogonal Latin hypercube ( nearly_orthogonal_lhs ) — minimizes correlation while retaining Latin hypercube stratification — @saudzahirr
Maximin and minimax distance designs ( maximin_design , minimax_design ) — spread points by optimizing pairwise or covering distances — @saudzahirr
Faure and Niederreiter low-discrepancy sequences ( faure_sequence , niederreiter_sequence ) — deterministic space-filling sequences for quasi-Monte Carlo sampling — @saudzahirr
Sequential Bayesian-optimization designs ( sequential_design ) with Gaussian-process regression and expected-improvement, probability-improvement, and upper-confidence-bound acquisition functions — @saudzahirr
feat: add Latin square family, Hammersley sequence, and Iman-Conover method by @saudzahirr in https://github.com/pydoe/pydoe/pull/128
Full Changelog: https://github.com/pydoe/pydoe/compare/v1.1.0...v1.2.0
Latin square ( latin_square ), Graeco-Latin square ( graeco_latin_square ), and hyper-Graeco-Latin square ( hyper_graeco_latin_square ) designs — remove the effect of two, three, or more nuisance factors using mutually orthogonal Latin squares — @saudzahirr
Hammersley point set ( hammersley_sequence ) — finite, fixed-size low-discrepancy point set with one equispaced coordinate and the remaining coordinates from the Halton sequence — @saudzahirr
Iman-Conover method ( iman_conover ) — induces a target rank correlation structure among sampled variables while preserving their marginal distributions — @saudzahirr
Mixture axial (screening) design ( mixture_axial_design ) — places one point at the simplex centroid and one axial point per component for screening blending effects — @saudzahirr
Extreme-vertices design ( extreme_vertices_design ) — enumerates the vertices of a constrained mixture region defined by per-component lower/upper bounds — @saudzahirr
Mixture-process variable design ( mixture_process_design ) — crosses a mixture design with an independent process-variable design — @saudzahirr
Blocking of full factorial designs ( block_full_factorial ) — splits a (2^k) factorial into (2^p) blocks by confounding chosen interactions with block effects — @saudzahirr
Orthogonally-blocked central composite design ( block_ccdesign ) — splits a CCD into a factorial block and an axial block with alpha chosen for orthogonal blocking — @saudzahirr
Hartley's small composite design ( small_composite_design ) — augments a resolution-III fractional factorial with star points, requiring fewer runs than a standard CCD — @saudzahirr
Definitive screening design ( definitive_screening_design ) — three-level Jones-Nachtsheim design requiring only (2k+1) runs, built from a Paley conference matrix and its fold-over — @saudzahirr
Supersaturated design ( supersaturated_design ) — random-search construction of (k > n) two-level designs minimizing (E(s^2)) for screening under effect sparsity — @saudzahirr
Nested Latin hypercube ( nested_lhs ) — constructs nested designs for multi-fidelity experiments — @saudzahirr
Orthogonal array-based Latin hypercube design ( oa_lhd ) — Tang's (1993) construction turning a symmetric orthogonal array into a Latin hypercube with improved two-dimensional uniformity — @saudzahirr
Sliced Latin hypercube design ( sliced_lhs ) — partitions an (N=mt) -point Latin hypercube into (t) slices of (m) points, each a Latin hypercube in its own right — @saudzahirr
Bump pymdown-extensions from 10.21.2 to 10.21.3 in the uv group across 1 directory by @dependabot[bot] in https://github.com/pydoe/pydoe/pull/116
Full Changelog: https://github.com/pydoe/pydoe/compare/v1.0.2...v1.1.0
Simplex-lattice design ( simplex_lattice_design ) — mixture experiment design placing lattice points on the q -component simplex at resolution 1/ m ; supports fitting Scheffé polynomial models of degree m — @saudzahirr
Simplex-centroid design ( simplex_centroid_design ) — mixture experiment design with the centroids of all (2^q - 1) non-empty subsets; supports estimation of all Scheffé interaction blending coefficients — @saudzahirr
John's 3/4 fractional factorial design ( john_three_quarter_design ) — semifoldover design using exactly 3/4 of the runs of the next full (2^k) design; de-aliases all two-factor interactions involving the chosen factor — @saudzahirr
Fix contributing docs to reference zensical instead of mkdocs serve — @saudzahirr
Add Mixture Designs reference page — @saudzahirr
Update factorial.md by @saudzahirr in https://github.com/pydoe/pydoe/pull/111
Full Changelog: https://github.com/pydoe/pydoe/compare/v1.0.1...v1.0.2
Removed deprecated code (v1.0.1) by @saudzahirr in https://github.com/pydoe/pydoe/pull/106
Full Changelog: https://github.com/pydoe/pydoe/compare/v1.0.0...v1.0.1
Basically no change since 0.9.9. This release 1.0 marks the completion of the transition from pyDOE3 to pydoe.
Basically no change since 0.9.9. This release 1.0 marks the completion of the transition from pyDOE3 to pydoe.
Full Changelog: https://github.com/pydoe/pydoe/compare/v0.9.9...v1.0.0
Update copyright info by @saudzahirr in https://github.com/pydoe/pydoe/pull/92
Full Changelog: https://github.com/pydoe/pydoe/compare/v0.9.8...v0.9.9
Migrate documentation tooling from MkDocs to Zensical ( #71 ) — @saudzahirr
Update copyright information ( #92 ) — @saudzahirr
Bump actions/download-artifact from 7 to 8 by @dependabot[bot] in https://github.com/pydoe/pydoe/pull/72
Full Changelog: https://github.com/pydoe/pydoe/compare/v0.9.7...v0.9.8
Bump mkdocs-material from 9.7.1 to 9.7.2 by @dependabot[bot] in https://github.com/pydoe/pydoe/pull/66
Full Changelog: https://github.com/pydoe/pydoe/compare/v0.9.6...v0.9.7
Update Artifact Name for Docs by @laraibg786 in https://github.com/pydoe/pydoe/pull/61
np.is_close for logical float comparison in pbdesign by @saudzahirr in https://github.com/pydoe/pydoe/pull/65Full Changelog: https://github.com/pydoe/pydoe/compare/v0.9.5...v0.9.6
Restructure codebase into modular package layout ( #63 ) — @laraibg786
Rename modules to snake_case to comply with PEP 8 ( #62 ) — @saudzahirr
Fix smoke test after publishing package by @laraibg786 in https://github.com/pydoe/pydoe/pull/60
Full Changelog: https://github.com/pydoe/pydoe/compare/v0.9.4...v0.9.5
Add pip ecosystem to dependabot for auto update of python dependencies by @laraibg786 in https://github.com/pydoe/pydoe/pull/39
Full Changelog: https://github.com/pydoe/pydoe/commits/v0.9.4
First public PyPI release.
Factorial designs resolution table in documentation ( #54 ) — @saudzahirr
Type annotations throughout the codebase ( #49 ) — @saudzahirr
Ruff as default linter and formatter with Codecov integration ( #47 ) — @saudzahirr
PyPI release workflow ( #57 ) — @laraibg786
uv as package manager with updated pyproject.toml ( #40 ) — @laraibg786
Dynamic versioning without a _version file ( #38 ) — @laraibg786
Buggy resolution values in fractional factorial designs ( #52 ) — @Noor-Mustafa123
lhsmu TypeError when using NumPy Generator API ( #44 ) — @saudzahirr
Negative determinant bug in d_efficiency ( #42 ) — @saudzahirr
Singular information matrix handling — @saudzahirr
Improve RAM usage in _lhsclassic ( #43 ) — @saudzahirr
Type annotations throughout the codebase — @saudzahirr
Type annotations throughout the codebase — @saudzahirr
Ruff as linter and formatter with Codecov integration — @saudzahirr
Migrate documentation site from Sphinx to MkDocs — @saudzahirr
Migrate documentation site from Sphinx to MkDocs — @saudzahirr
Integrate uv as package manager — @laraibg786
Dynamic versioning without a _version file — @laraibg786
Fix lhsmu TypeError when using NumPy Generator API — @saudzahirr
Fix negative determinant bug in d_efficiency — @saudzahirr
Improve RAM usage in _lhsclassic — @saudzahirr
Fix singular information matrix handling — @saudzahirr
Add doe_sparse_grid using Smolyak's sparse grid construction — @saudzahirr
Replace legacy NumPy RandomState with Generator API throughout — @saudzahirr
Deprecate random_state parameter in lhs in favour of seed — @saudzahirr
Add morris_sampling for Morris one-at-a-time sensitivity screening — @saudzahirr
Add saltelli_sampling for variance-based Sobol' sensitivity analysis — @saudzahirr
Add random_uniform and random_k_means space-filling designs — @saudzahirr
Deprecate random_state parameter in lhs in favour of seed — @saudzahirr
Add halton_sequence , sobol_sequence , korobov_sequence , rank1_lattice , and sukharev_grid — @saudzahirr
Add halton_sequence , sobol_sequence , korobov_sequence , rank1_lattice , and sukharev_grid — @saudzahirr
Add cranley_patterson_shift for randomising quasi-random sequences — @saudzahirr
Add optimal_design with D, A, E, G, I, V, C, S, T optimality criteria — @saudzahirr
Add optimal_design with D, A, E, G, I, V, C, S, T optimality criteria — @saudzahirr
Add Fedorov, modified Fedorov, DETMAX, and Dykstra algorithms — @saudzahirr
Add doehlert_shell_design and doehlert_simplex_design — @saudzahirr
Add taguchi_design with orthogonal array lookup, SNR computation, and TaguchiObjective enum — @saudzahirr
Update gsd to accept level count of 1 — @saudzahirr
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Migrate build system to pyproject.toml with Hatchling — @saudzahirr
Migrate build system to pyproject.toml with Hatchling — @saudzahirr
Add full test suite and CI automation — @saudzahirr
Replace Black with Ruff; fix fracfact_opt ; add doc-tests — @saudzahirr
Refactor fracfact with test updates and minor bug fixes — @saudzahirr
Add gsd (Generalized Subset Design) — @saudzahirr
Fix integer division for Python 3 compatibility — @tisimst
Fix incorrect indexing variable in _pdist — @tisimst
Clarify BSD 3-clause license — @tisimst
Add code credits to original authors — @tisimst
Add var_regression_matrix to public API — @tisimst
Fix Python 3 import and compatibility issues — @tisimst
Remove package import for better Python 3 compatibility — @tisimst
Fix indexing problem; improve Python 3 compatibility — @tisimst
### Fixed - Fix indexing issue — @tisimst
Simplify design function inputs to accept number of factors directly — @tisimst
Simplify design function inputs to accept number of factors directly — @tisimst
Improve consistency across all design functions — @tisimst
Add criterion keyword argument to lhs for more flexible sampling — @tisimst
Fix syntax errors and update documentation links — @tisimst
Add response surface designs ( bbdesign , ccdesign ) — @tisimst
Add response surface designs ( bbdesign , ccdesign ) — @tisimst
Add Latin hypercube sampling ( lhs ) — @tisimst
Initial release with fullfact , ff2n , fracfact , pbdesign , and gsd . — @tisimst
Initial release with fullfact , ff2n , fracfact , pbdesign , and gsd . — @tisimst
Back to top
One column per quarter.