迭代遗传算法及其用于生物反应器补料优化

针对化工动态优化的数值求解问题，提出将迭代思想与遗传操作相结合，构建迭代遗传算法.算法首先对时间区间和控制搜索域实施离散化，进而应用遗传操作搜索离散问题的最优控制策略.逐步收缩搜索域并迭代以消减离散化带来的偏差，不断改善寻优结果，增强算法的稳健性.实例测试表明该算法简便、可行、高效，已成功地应用于Lee-Ramirez生物反应器补料流率的优化，运算结果优于文献值，显示了迭代遗传算法的优越性.迭代遗传算法尤其适用于系统的梯度信息不可得的情况.

Iterative genetic algorithm and its application to feed policies optimization for bioreactor

To solve dynamic optimization problems of chemical processes with numerical methods, a new algorithm,iterative genetic algorithm (IGA) was developed, of which the main idea was to iteratively call genetic operations and gradually approximate the optimal control profile.The first step of IGA was to discretize time interval and control region to make the continuous dynamic optimization problem a discrete problem.Genetic operations were then used to seek the best control profile of the discrete dynamic system.Lastly,region-reduction strategy and iteration were used to increases numerical accuracy and enhance robustness.IGA was shown by a case study to be convenient, feasible, and efficient.When applied to optimizing feed-rate of Lee-Ramirez bioreactor, IGA displayed advantage over general methods.The result by IGA was better than that in the reference.IGA approach could be regarded as a reliable and useful optimization tool when the gradients are not available.