基于两级参数算法的对偶控制

Dual control based on two-level parametric algorithm

针对具有未知参数的随机系统,从随机次优的角度出发,将原不可解的动态规划问题转化为优化一个效用函数,该效用函数考虑了输出调节要求和参数学习要求及两者之间的折衷,充分利用了对偶控制较常规自适应控制的优越性.提出利用两级参数算法来最小化效用函数,从而获得控制信号.仿真结果表明,这种控制器具有良好的对偶性质,能得到较好的学习和控制效果.

A dual control method is proposed for stochastic systems with unknown parameters. The utility function to be minimized indicates the desired degree of compromise between the regulate effect of the output and the learning effect of the unknown parameters, taking advantage of the superior performance of dual control compared to conventional adaptive control. The two-level parametric algorithm is introduced to derive the control law. Simulation results show the dual property of the controller that can achieve good estimation results as well as good control performance.