Optimal experimental design using partial least squares regression

Optimal experimental design is used to either reduce the number of experiments to be performed or to get as much information from the available data and data retrieved from a pre-determined set of experiments. An optimal experimental design aims at minimizing the cost of experimentation and still getting useful data to determine the process properties. This data in turn can be used to construct a black box model of the process response with respect to the input factors and use that model to find an optimum operating regime to get a desired response from the system. This work presents a procedure using partial least squares regression which can be used to design an experiment and iteratively improve the model using the data from the designed experiment. The technique is illustrated with a case study of membrane ultrafiltration process for separating hemoglobin from bovine serum albumin.