Advanced modelling for media composition optimization of phototrophic growth

An Artificial Neural Network model was implemented in order to correlate media composition with the maximal efficiency in biomass and the maximal specific growth rate of the purple bacteria Rhodopseudomonas palustris. The performance of the ANN model was compared with the full quadratic model used in the response surface methodology to optimize the media composition. The fit grade was improved for both response variables by the ANN, being particularly effective for the specific growth rate, which was improved up to 20 percentage units. © 2010 IFAC