Parameter Estimation of Biosurfactant Production from Agro-industrial Waste Using Genetic Algorithm

Although various kinds of kinetic models were proposed to describe the dynamic behaviour of biologically reacting systems, very little studies have been carried about biosurfactant formation. Its sustainable production, in renewable culture medium, is even more complex. In this sense, emerges the Genetic Algorithm (GA), an effective stochastic global search algorithm inspired in the evolution theory that has been showing a great potential to find optimal solutions in complex systems. This work aims to evaluate the GA's estimation of parameters involved in biosurfactant production from agro-industrial waste using Bacillus subtilis. Three different models were proposed to describe biomass growth, substrate consumption, biosurfactant synthesis and dissolved oxygen in the medium. The technique's quality was evaluated by the sum of squared errors (SSE) and correlation coefficient (R-2). The results indicated that the best model to describe the system's dynamics obtained SSE lower than 1 and R-2 superior to 0.97 for almost all the variables.