基于蚁群算法的模糊控制规则的过滤简化

针对设计高维模糊控制器过程中会遇到的“规则爆炸”问题，利用蚁群算法进行控制规则的过滤简化。为了用尽量少的规则得到尽可能好的控制效果，利用蚁群算法在饵决组合优化问题中的强大优势，在已有的完备规则中优选出若干条规则嵌人模糊控制器。采用带有时间窗口的蚁群算法去克服遗传算法优选模糊控制规则时可能产生的规则不连续的问题。该文还从遗传算法和蚁群算法工作机制的角度分析了对这两种算法加入约束条件的可操作性。以单级倒立摆控制系统为对象进行仿真研究，最后的仿真结果表明该文方法可以使模糊控制规则具有更好的简化效果和鲁棒性，并能具有好的控制效果。

Filtering Fuzzy Control Rules Based on Ant Colony Algorithm

To solve the typical "fuzzy rule explosion" problem in designing the high dimensional fuzzy controller, ant colony algorithm is used to filter a designed complete fuzzy-rule base. In order to get better control results with less control rules, ant colony algorithm is used, because of its superiority in solving combination optimization problems, to select several "good" fuzzy rules from a complete fuzzy rule base which is used to construct the Fuzzy controller. Ant colony algorithm with "time window" is used to overcome the problem of getting discontinuous fuzzy rules when selecting fuzzy rules with genetic algorithm. This article also analyzes the maneuverability of adding restriction in the two algorithms (GA and AS) from the mechanism point of view. At last, inverted pendulum is used as controlled plant to do the simulation. Results show that this method is effective enough to make the fuzzy controller much simpler and robust and to get good control performance.