Abstract: The multi-output support vector regression with composite kernel and the fuzzy theory were applied to design permanent magnet drive. In this method， the space particle swarm optimization (SPSO) algorithm was firstly introduced to obtain the most appropriate parameter of the multi-output support vector regression with composite kernel model. In addition， through the experiment the regression model between performances and structure parameters of permanent magnet drive was established. Secondly， by using fuzzy theory， multi-objective problem was converted into single one， and the mathematical model of optimization problem was set up， which was solved by SPSO. Finally， precision analysis of model， ANSYS simulation and prototyping test were carried out， and the results verified the effectiveness of the proposed method.

Key words: multi-output support vector, fuzzy theory, space particle swarm optimization algorithm, permanent magnet drive, multi-objective design