Abstract: A new method was researched for system model identification via Particle Swarm Optimization（PSO） in this paper.The basic idea of the method employs a system model composed with classical models so as to transform the system structure identification problem into a combinational problem.The PSO is then adopted to implement the identification on the system structure and parameters.In order to enhance the performance of the PSO identification，a Velocity Mutation Particle Swarm Optimization（VMPSO） algorithm is also presented.Finally，simulation results indicate the rationality of the novel identification method and the effectiveness of the solving algorithm.

Key words: system identification, particle swarm optimization, velocity mutation, meta model, combinatorial optimization

摘要： 论文研究了一种利用粒子群优化（PSO）算法对系统模型进行辨识的新方法。该方法的基本思想是将典型的数学模型的相互组合而构成系统模型的新颖辨识方法，即首先将系统结构辨识问题转化为组合优化问题，然后采用粒子群优化算法同时实现了系统的结构辨识与参数辨识。为了进一步增强粒子群优化算法的辨识性能，提出了一种利用速度变异的粒子群优化（VMPSO）算法。最后，给出了仿真示例，其结果表明了所给的系统辨识方法的合理性和求解算法的有效性。

关键词: 系统辨识, 粒子群优化算法, 速度变异, 元模型, 组合优化