A. Abakarov

MULTIOBJECTIVE OPTIMIZATION APPROACH: THERMAL FOOD PROCESSING

The objective of this study was to utilize a multiobjective optimization technique for the thermal sterilization of packaged foods. The multiobjective optimization approach used in this study is based on the optimization of well-known aggregating functions by an adaptive random search algorithm. The applicability of the proposed approach was illustrated by solving widely used multiobjective test problems taken from the literature. The numerical results obtained for the multiobjective test problems and for the thermal processing problem show that the proposed approach can be effectively used for solving multiobjective optimization problems arising in the food engineering field.

Estimating Reaction Rates in Squid Protein Hydrolysis Using Artificial Neural Networks

In this study, the use of artificial neural networks (ANN) for estimating reaction rates in enzymatic hydrolysis of squid waste protein was investigated. This is a complex process because a number of inherent simultaneous inhibition and enzyme inactivation reactions occur during hydrolysis that make it difficult to develop a reliable kinetic model by more traditional deterministic approaches. A series of 12 enzyme hydrolysis experiments were carried out on samples of squid waste under specified conditions of temperature, pH, and initial enzyme and substrate concentrations. Experimental data in the form of substrate concentration over time were taken as real time course data (TCD), and divided into three groups for respective use in training, validating, and testing the model. A feed-forward architecture was utilized to construct the necessary predictive model. The network was trained until the mean squared error function between target and actual output values reached a desired minimum. Data sets from the remaining two groups were used for subsequent validation and testing of the model. The model performed well when tested against experimental data in the third group (not used in its development) and taken over a wide range of initial conditions. Maximum differences between experimental and predicted values of substrate concentration at any point in time ranged from 0.3 to 0.5 g L-1 (1% to 3% of initial substrate concentration), with correlation coefficients between predicted and experimental results ranging from 0.95 to 0.97.