永磁球形电机的支持向量机模型的参数寻优

Parameter Optimization for Support Vector Machine Model of Permanent Magnet Spherical Motors

提出一种新的永磁球形电机的数学模型建立方法——支持向量机(SVM)模型,分别采用网格参数寻优、遗传算法(GA)参数寻优、粒子群(PSO)寻优算法来计算SVM模型的参数,分析比较三种算法得到的参数,确定最适合永磁球形电机SVM回归模型的参数寻优算法。

A new mathematical modeling method named support vector machine（SVM） model for permanent magnet spherical motors is presented. Three parameter optimization algorithms based on grid search, genetic algorithm（GA） and particle swarm optimization（PSO） are used respectively. The parameters obtained by three algorithms are analyzed and compared, and the algorithm which most fits the SVM model of permanent magnet spherical motor is confirmed.