ANN AND RSM BASED OPTIMIZATION OF CELLULASE PRODUCTION BY HYPOCREA SP Z28 BY SUBMERGED FERMENTATION

Response surface methodology (RSM) and artificial neural network (ANN) were used to simulate and optimize cellulase production by Hypocrea sp. Z28 by submerged fermentation. Results showed ANN had higher simulation accuracy than RSM. Cellulase production optimized by RSM was 5.48 U/mL, while the corresponding experimental value was 5.67 +/- 0.32 U/mL. Using ANN as a prediction function, a maximum cellulase production of 5.96 U/mL was searched by the genetic algorithm, and the corresponding experimental value was 6.01 +/- 0.43 U/mL. Compared to RSM, ANN brought higher and more accurate cellulase production values. The application of ANN to optimize cellulase production proved successful