基于遗传算法的LQ控制器的权值优化

摘要

LQ(LQR & LQG)is the most widely used algorithm of an optimal control theory applied to industrial control.Its answer can be got from the solution of Riccati equation.And the choice of the two weight matrixes (Q & R)influences greatly whether the solution of the equation is effectual.How to choose Q and R to fulfil the optimal control is a very important thing for all process control engineers.In this paper,the author uses genetic algorithm to find the the global optimal anawer.The simulated results are satisfactory.%线性二次型最优调节器及线性二次型高斯问题，通称LQ问题，是最优控制在民用工业中应用最广的算法。LQ问题的解，最终归结为求解Riccati方程，而方程解的有效性，很大程度上又依赖于两个权矩阵Q阵和R阵的选择，如何选择Q阵和R阵以保证实现控制目标最优一直为控制界所注。本文提出以遗传算法来达到实现全局寻优的目的，得到较满意的结果。