Modeling of Chromatography allows a better understanding and development of new techniques to be applied at industrial level, although it's relatively complex. The models of this process are represented by systems of partial differential equations with non linear parameters difficult to estimate generally, which constitutes an inverse problem. In general there aren't analytical solutions and therefore numerical methods should be used for their direct solutions. Frequently typical boundary conditions are considered, but it's convenient to study different approaches for those.
Evolutionary Computation has been used successfully in many problems of diverse areas for searching in complex spaces. Considering previous works from the authors, in this article Genetic algorithm and Differential evolution are used for parameters estimation in models of protein chromatography with variants in boundary conditions. In both algorithms each population individual is a supposed condition to the direct solution for the system of partial differential equations, coded in real values, while inverse solution is optimized updating the first one according to a fitness function. A comparative analysis is showed as result.
Cite This Article
Irízar, M., Câmara, L. D., J., A., Llanes, O. (2009). Inverse Solution of a Chromatography Model by means of Evolutionary Computation.
CMES-Computer Modeling in Engineering & Sciences, 54(1), 1–14. https://doi.org/10.3970/cmes.2009.054.001