基于混沌和SVM的青霉素发酵过程混合建模  

Hybrid modelling of penicillin fermentation process based on chaos and SVM

作　　者：王鲜芳[1,2] 潘丰[1] 

机构地区：[1]江南大学通信与控制工程学院自动化研究所,江苏无锡214122 [2]河南科技学院信息工程系,河南新乡453003

出　　处：《计算机与应用化学》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)

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.

关 键 词：混沌算法 支持向量机 状态变量 智能控制 

分 类 号：TQ015.9[化学工程] O6-39[理学—化学]