In many experiments, the same experimental unit is evaluated with respect to more than one response. Such multiresponse experiments are optimized by searching for a point in the design region where responses best satisfy a set of criteria. Typical optimization criteria are minimizing distance-to-target and variance. This article explores minimizing the prediction variance as a criterion in the multiresponse optimization on a case. In the case, selected from the food processing industry, we wish to optimize a military food product where 24 sensory attributes are evaluated in a sensory panel. The prediction variance captures the variability in the responses adjusted for how well we can predict them, given the experiment performed to estimate the parameters in the prediction model. Because the underlying regression models for the responses are critical in estimating the prediction variance, this article presents a tutorial on various models for fitting models to multiple responses.
All Science Journal Classification (ASJC) codes
- Safety, Risk, Reliability and Quality
- Industrial and Manufacturing Engineering
- Military food product
- Multiresponse optimization
- Prediction variance