The mathematical modeling for the optimization of triacyl glycerol acylhydrolase production through artificial neural network and genetic algorithm

Abstract

The demand of industrial enzymes is growing tremendously worldwide due to the intervention of enzyme in the various commercial and industrial applications. The fulfillment of the market demand, several production strategies have been employed by researcher to get the optimal yield. Mathematical modeling helps in deriving a non-linear equation by considering the individual, square and interaction effects of the process variables on the product formation.  The soft computing-based optimization helps in attaining the optimum output of the production without trapping in local optima. The influencing parameters for enzyme productions considered in this research work are temperature (°C), liquid to solid ratio, pH and incubation time (hours). Artificial Neural Networks (ANN) and Genetic Algorithms (GA) serve as a better modeling tool and optimization approach for enzyme production due to the better search and optimal criteria. In this research work with mathematical modeling for the optimization of triacylglycerol acylhydrolase production using agro-residue as a substrate was carried out using Artificial Neural Network (ANN) and Genetic Algorithm (GA). The Feed-Forward-Back Propagation (FFBP) algorithm along with CCD data was used for the ANN model development. This model also supported with regression plots which show the high regression coefficient (R² =0.75). The optimization of this model shows that the enzyme production can be improved up to ~88.52 % at temperature 35°C, liquid-solid ratio is 1.5, pH 7 and incubation time 120 h. The proposed optimized model helps in scale-up studies of enzyme production without any difficulty. This research work revealed that the enhanced process attributes of ANN-GA over OVAT (One Variable at a Time) approach.