ForgeDOE

Design of Experiments engine. Classical, optimal, screening, adaptive, and space-filling designs with full analysis pipeline.

Install

pip install forgedoe

Quick Start

from forgedoe.core.types import Factor
from forgedoe.designs.factorial import full_factorial
from forgedoe.analysis.regression import fit_model

factors = [Factor("Temp", 150, 200), Factor("Time", 10, 30)]
design = full_factorial(factors)
result = fit_model(design, responses=[85, 90, 78, 95])

Adaptive Bayesian DOE

from forgedoe.adaptive.bayesian_doe import AdaptiveExperiment

exp = AdaptiveExperiment(factors, n_initial=6)
while not exp.should_stop():
    point = exp.suggest_next_point()
    exp.add_observation(point, measured_value)

Modules

Module	Contents
core.types	Factor, Response, DesignMatrix, AnalysisResult
core.coding	Coded/natural unit encoding
designs.factorial	Full factorial, fractional factorial, Plackett-Burman
designs.response_surface	Central composite (CCD), Box-Behnken
designs.screening	Definitive screening designs (DSD)
designs.space_filling	Latin hypercube, maximin LHS
analysis.regression	Model fitting, ANOVA, effects, diagnostics
analysis.optimization	Desirability functions, multi-response optimization
adaptive.bayesian_doe	Bayesian adaptive experimentation
adaptive.sequential	Sequential experiment planning
power.power_analysis	Power analysis, sample size, required replicates
calibration	Self-calibration against golden references

Dependencies

• numpy
• scipy

License

MIT