基于人工鱼群算法的随机结构AMD控制系统LQR权矩阵优化

针对随机结构在平稳随机激励下线性二次最优控制的权参数选取问题，提出了基于人工鱼群算法的随机结构AMD控制系统权矩阵优化设计方法。该方法以结构随机响应和控制力均方值的加权组合为目标函数，考虑了结构和外激励的随机性对于控制效果的影响。其优化结果不仅保证了控制器的控制效果，而且降低了控制效果对于随机参数的敏感性，增强了控制器的鲁棒性。最后结合数值算例，验证了所述方法的有效性和正确性。

Optimal analysis of weight matrices of LQR algorithm for stochastic structure-AMD system based on artificial fish algorithm

In allusion to the choice of the weight matrices Q and R in LQR algorithm in the stochastic structure-AMD system under the stationary random excitation, the optimization method of the weight matrices is proposed based on the artificial fish algorithm. The method resolves the weight matrices optimization via the objective function with the weighted root mean square of the stochastic structural responses and active control force, taking into account the effect of the randomness in the structure and excitation on the control effect. The effectiveness and validity of the procedures of the optimal designs of the weight matrices is proved by a numerical example. Numerical results preliminarily show that the optimized weight matrices can effectively suppress the undesirable vibration, reduce the sensitivity of the control effectiveness to the random parameters and improve the robustness of AMD under the earthquake.