论文部分内容阅读
就青霉素发酵过程难以建立理想模型,提出一种基于混沌支持向量机和动力学模型相结合的混合建模新方法。首先分析青霉素发酵过程动力学模型的特点,选择合适的状态变量,然后利用混沌算法优化支持向量机的参数,建立动态时变的混合模型。该模型不但能自动选择支持向量机的参数,而且能够预报一些不能在线测量的生化状态变量。通过实用,证明了此方法有效。
It is difficult to establish an ideal model for penicillin fermentation process. A new hybrid modeling method based on chaos support vector machine and dynamic model is proposed. Firstly, the characteristics of kinetic model of penicillin fermentation process were analyzed, the appropriate state variables were selected, and then chaos algorithm was used to optimize the parameters of support vector machine to establish a dynamic time-varying hybrid model. The model can not only automatically select SVM parameters, but also can predict some biochemical state variables that can not be measured online. Through practical application, this method is proved to be effective.