Recurrent Neural Networks and Micro-Genetic Algorithms for Estimation and Optimization: Application to Bioprocesses

该项目旨在通过提供一种在线过程变量估计和优化的新技术来帮助生物过程监测、控制和优化。发酵动力学模型很少考虑环境因素的影响，如pH和温度。开发精确的定量模型，模拟细胞如何对各种环境变化做出反应，将有助于更好地利用细胞的化学合成能力。提出的方法将使用在线传感器提供的环境和生理信息来估计生物反应器中物种的浓度。在这些估计的基础上，优化方案将提供最佳的环境条件，使电池达到最大生产率。

This project is aimed at helping bioprocess monitoring, control, and optimization by providing a novel technique for on-line process variable estimation and optimization. Fermentation kinetic models seldom include the influence of environmental factors, like pH and temperature. The development of accurate quantitative models, which simulate how cells respond to various environmental changes, will help in better utilizing the chemical synthesis capabilities of cells. The proposed method will use environmental and physiological information available from on-line sensors to estimate concentration of species in a bioreactor. Based on these estimates, the optimization scheme will provide the best environmental conditions which will drive the cells to their maximum productivity.