Product Optimization using Monte Carlo Simulation to Incorporate Factor Variation

Experimental designs play a crucial role in product/process characterization.  However, the model built from the experiment may not provide the best optimization because it only considers uncertainty through the model error.  Simulation can be used to incorporate the uncertainty in the factors allowing the experimenter to better estimate the true variation in the response. This provides a more realistic estimate of product/process capability. The method is illustrated using a fermentation example with a single response and several input factors. Considerations for multiple responses will be discussed.