Optimization on response surface models for the optimal manufacturing conditions of dairy tofu

Abstract

The purpose of this research was to create response surface models through regression on experimental data and to apply the Sequential Quadratic Programming (SQP) and Genetic Algorithms (GAs) on the models to obtain optimal processing conditions for dairy tofu. The two-stage effort of obtaining a surface model using response surface methodology (RSM), and optimizing this model using GAs or SQP techniques was demonstrated to be an effective approach. Both SQP and GAs techniques were able to determine the optimal conditions for manufacturing the probiotic dairy tofu. The conditions were 1% of glucono-delta-lactone (GDL), 0% of peptides level, 3% of isomaltooligosaccharides (IMO) and 18% of milk concentrations, and they were confirmed by verification experiments. Among the SQP and two GAs employed, the SQP, modified with the multi-start capability, is the most efficient one.