Model penicillin fermentation by least squares support vector machine with tuning based on amended harmony search  

作　　者：Hai-Bin Ou Yang Steven Li Ping Zhang 

机构地区：[1]College of Information Science and Engineering Northeastern University Shenyang 110819, P. R. China [2]Graduate School of Business and LawRMIT University, Melbourne 3000, Australia [3]Anshan Normal University, Anshan Liaoning 114005, P. R. China

出　　处：《International Journal of Biomathematics》2015年第3期175-204,共30页生物数学学报（英文版）

基　　金：The authors wish to thank the editor and anonymous referees for their constructive comments and recommendations, which have significantly improved the presentation of this paper. This work is supported by National Nature Science Foundation of China （Grant Nos. 60674021, 61273155）.

摘　　要：Penicillin fermentation is an important part of microbial fermentation. Due to the existence of error date in the independent variables and dependent variables of the penicillin fermentation sample data, the accuracy of the model of penicillin fermentation is affected. In this paper, an amended harmony search （AHS） algorithm is developed to adjust the hyper-parameters of least squares support vector machine （LS-SVM） in order to build penicillin fermentation process model with prediction accuracy. The AHS algorithm is investigated by unconstrained benchmark functions with different characteristics. Compared with other several optimization approaches, AHS demonstrates a better performance. Moreover, using the simulation data from the PenSim simulation platform to validate the effectiveness of the penicillin fermentation process modeling, experiment results show that the penicillin fermentation process modeling based on the tuned LS-SVM by AHS possesses robustness and generalization ability.

关 键 词：Penicillin fermentation harmony search algorithm LS-SVM penicillin fermentation process model robustness. 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置] TP18[自动化与计算机技术—控制科学与工程]