[1]WANG Jie,CHEN Kai-peng.Application Study of SVM Predictive Control Based on DecisionFunctions Simplification and Pso Optimization[J].Journal of Zhengzhou University (Engineering Science),2013,34(02):53-56.[doi:10.3969/j.issn.1671-6833.2013.02.014]
Copy
Journal of Zhengzhou University (Engineering Science)[ISSN
1671-6833/CN
41-1339/T] Volume:
34
Number of periods:
2013 02
Page number:
53-56
Column:
Public date:
2013-03-28
- Title:
-
Application Study of SVM Predictive Control Based on DecisionFunctions Simplification and Pso Optimization
- Author(s):
-
WANG Jie; CHEN Kai-peng
-
School of Electrical Engineering,Zhengzhou University,Zhengzhou 450001,China
-
- Keywords:
-
SVM; decision function; PSO; predictive control
- CLC:
-
TP181
- DOI:
-
10.3969/j.issn.1671-6833.2013.02.014
- Abstract:
-
For large-scale samples,SVM does not perform as well as neural networks and this paper tries toimprove the training speed by simplify the decision functions by matrix transform. We simplify the unnecessarysupport vectors in SVM modeling and introduce a relaxation factor in order to improve the effects of simplifica-tion.Experiment shows that the number of support vectors is redused by at least one third. Using the modelnonlinear model built by SVM after linearizing as the predictive model of predictive control. PSO was used toselect the best SVM parameters and computing the optimal control law of predictive control. The method canaccelerate the response and shorten the overshoot through a simulation of a cement rotary kiln.