中文摘要：

针对常规的自适应动态矩阵控制算法,在实际被控对象的模型参数发生突变时,系统的瞬态响应较差,提出了基于PSO在线优化的多模型自适应动态矩阵控制方法。对一类含跳变参数的单输入单输出离散时间被控对象,在模型参数范围未知情况下,以自适应模型参数为依据,经规则判断后通过所提出的基于双群体深度搜索的粒子群优化（PSO）算法在线优化自适应模型参数,并通过所定义的模型相似度自动建立多个固定模型。通过指标切换函数找到当前最优控制器。仿真结果表明,该方法明显优于常规的自适应动态矩阵控制算法,说明了该方法的有效性和可行性。

英文摘要：

Focusing on the traditional adaptive dynamic matrix control, when the parameters of the virtual model varies, the system response turns bad, a multiple models adaptive dynamic matrix control algorithm based on PSO online optimization is proposed. Applying the particle swarm optimization （PSO） algorithm based on double groups further searching, adaptive model parameters is optimized and fixed models are built. The current optimum controller is found due to switching function. Simulation results show its efficiency and availability, and this method is better than the traditional one.