摘要
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。
A new hybrid modeling method based on chaotic support vector machines and kinetics model is proposed to solve hard mod- eling in penicillin fermentation process. The first step is analyzing the characteristics of the kinetics model for Penicillin fermentation process, the second is choosing some suitable state variables, the third is was optimizing the parameters of support vector machines with Chaos algorithm, and the last is building a time-varying dynamic hybrid model. The model not only could realize choosing parameters of SVM automatically, but also realize online pre-estimation for some biochemical state variables which could not be on-line measured. It has been proved that the method is efficient through the practical application for the fermentation process.
出处
《计算机与应用化学》
CAS
CSCD
北大核心
2009年第4期451-454,共4页
Computers and Applied Chemistry
基金
国家高技术研究发展计划(863计划)重点项目(2006AA020301)
关键词
混沌算法
支持向量机
状态变量
智能控制
chaos algorithm, support vector machine, state variables, intelligence control
