Abstract: A Dual control algorithm of Generalized Predictive Adaptive Control（DGPAC） is presented，it considers errors of estimated parameters sufficiently in the process of identifying model，accumulates the information of estimated parameters maximal at the same time of system states optimum moving along the scheduled track so as to quickly reduce uncertainties.The simulation results show that this control algorithm is more effective than common generalized predictive adaptive control.

Key words: generalized predictive adaptive control, dual control, Controlled Auto-Regressive Integrated Moving Average（CARIMA） models, parameter estimation

摘要： 提出了一种基于双重控制思想的广义预测自适应控制算法，该算法在模型辨识和控制的过程中，采用谨慎控制和探测控制相结合的双重控制，充分考虑估计参数的误差，在使系统状态最优地沿预定轨线运动的同时最大限度的积累被估计参数的信息，以便最快地降低系统的不确定性。仿真结果表明，该控制算法比普通的广义预测自适应控制具有更好的控制品质。

关键词: 广义预测自适应控制, 双重控制, 受控自回归滑动平均模型, 参数估计