基于粗糙集与PSO-SVM的二甲醚催化燃烧预测模型

Prediction Model of Dimethyl Ether Catalytic Combustion Based on Rough Set and PSO- SVM

影响二甲醚催化燃烧的因素较多,并且影响因素之间又相互作用,将支持向量机引入二甲醚预测领域.为了克服LS-SVM的局限性,提出将粗糙集理论、粒子群优化算法以及SVM算法相结合,对LS-SVM的惩罚因子和径向基函数的参数进行优化选取.对二甲醚催化燃烧进行了实例研究.结果表明,所提出的模型提高了网络的预测精度,验证了该模型的有效性.

In view of many factors that affect the dimethyl ether catalytic combustion, and there are the interaction between influence factors, the characteristics of the support vector machine （SVM） is introduced into dimethyl ether prediction fields. In order to overcome the limitations of the LS - SVM, the rough set theory and particle swarm optimization algorithm is proposed to combine the SVM algorithm, the optimization selection is carried on for the KS - SVM penalty factor C and radial basis function δ parameters. For dimethyl ether catalytic combustion is carried on the case study. The results show that the proposed model improves the prediction precision of the network, verifies the validity of the model.