INVESTIGATION OF THE STATISTICAL MODEL BASED OPTIMIZATION ACCURACY UNDER EXPERIMENTAL ERRORS. CASE STUDY: OPTIMIZATION OF NUTRIENT MEDIA FOR MICROORGANISMS’ CULTIVATION PROCESS

Abstract

The response surface methodology based process optimization procedure, including design of factorial experiments, development of statistical model and estimation of the optimum response point is simulated in order to investigate influence of experimental errors and experimental design area on the optimization accuracy. The investigated practical problem is optimization of nutrient media composition for cultivation of microorganisms’ culture. The optimization results under a priori estimated experimental errors and various factor variation ranges in factorial experiments are investigated and the relationships between the factors variation range and the confidence interval of the optimum point estimations are determined.