Design of Experiments (DOE)

Planning experiments with systematic data collection

Passive data collection leads to a number of problems in statistical modeling. Observed changes in a response variable may be correlated with, but not caused by, observed changes in individual factors (process variables). Simultaneous changes in multiple factors may produce interactions that are difficult to separate into individual effects. Observations may be dependent, while a model of the data considers them to be independent.

Designed experiments address these problems. In a designed experiment, the data-producing process is actively manipulated to improve the quality of information and to eliminate redundant data. A common goal of all experimental designs is to collect data as parsimoniously as possible while providing sufficient information to accurately estimate model parameters.

Functions

expand all

DOE Objects

`fullFactorialDOE`	Full factorial design of experiments (DOE) (Since R2024b)
`mixtureDOE`	Design of experiments (DOE) for mixture experiments (Since R2024b)
`optimalDOE`	D-optimal design of experiments (DOE) (Since R2024b)

Work with DOE Objects

`fitlm`	Fit linear regression model using design points (Since R2024b)
`addruns`	Add runs to D-optimal design (Since R2024b)

Full Factorial Designs

`ff2n`	Two-level full factorial design
`fullfact`	Full factorial design

Fractional Factorial Designs

`fracfact`	Fractional factorial design
`fracfactgen`	Two-level fractional factorial design generators

Response Surface Designs

`bbdesign`	Box-Behnken design
`ccdesign`	Central composite design

D-Optimal Designs

`candexch`	D-optimal design from candidate set using row exchanges
`candgen`	Candidate set generation
`cordexch`	Coordinate-exchange D-optimal design
`daugment`	D-optimal augmentation
`dcovary`	D-optimal design with fixed covariates
`rowexch`	Row exchange
`rsmdemo`	Interactive response surface demonstration

Latin Hypercube Designs

`lhsdesign`	Latin hypercube sample
`lhsnorm`	Latin hypercube sample from multivariate normal distribution

Quasi-Random Designs

`haltonset`	Halton quasirandom point set
`qrandstream`	Quasirandom number stream
`sobolset`	Sobol quasirandom point set

DOE Plots

`interactionplot`	Interaction plot for grouped data
`maineffectsplot`	Main effects plot for grouped data
`multivarichart`	Multivari chart for grouped data
`rsmdemo`	Interactive response surface demonstration
`rstool`	Interactive response surface modeling

Topics

Full Factorial Designs
Designs for all treatments
Fractional Factorial Designs
Designs for selected treatments
Response Surface Designs
Quadratic polynomial models
Improve an Engine Cooling Fan Using Design for Six Sigma Techniques
This example shows how to improve the performance of an engine cooling fan through a Design for Six Sigma approach using Define, Measure, Analyze, Improve, and Control.
D-Optimal Designs
Minimum variance parameter estimates