Genetic algorithm optimized artificial neural network - co-optimization
Description
Co-optimization of two response biomass (X), and PHA (P) based on RSM design - modeling and optimization. The second/ independent path takes the ANN approach. The same experimental design was exploited to have ANN-based modeling - ANN structure optimized with GA. The ANN model has been further utilized to co-optimize the responses.
Files
Steps to reproduce
The file "DOE_analysis_Coded_values" uses the experimental data set "CCD_bio_PHA_03022022" to model and optimize the process. The ANN approach "ANN1_GA_optimizing_Hyper_para" makes the ANN model finetunes hyperparameters, saves the network, bias, etc. "ANN2_optimal_Hyper" just validates the result assessed in the previous script. "ANN3_Surface_ plot" of the ANN model. "ANN4_GA_optimizing_Biomass_PHA" is the co-optimization of the responses using the ANN model under the GA framework and desirability.