本論文中，我們將討論動態灰色預測控制器的設計。我們嘗試著結合模糊理論、灰色預測控制以及基因演算法之優點來發展這樣的動態控制系統。我們將注意力集中在預測模式的變化轉換，不同於傳統方法以動態的預測步距選取為主。預測模式間的轉換及時機的規劃是我們設計的出發點，而且我們也會探討並非整個控制過程中都適合以預測的行為來控制系統，所以我們嘗試著去改變調整完整的控制器架構。最後由基因演算法來搜尋複雜的系統參數組合。藉由模擬的範例結果，我們可以發現所提出方法能使響應可以追蹤所期望的較佳之響應軌跡。 In this thesis, we will discuss the dynamic grey prediction systems. We try to integrate the advantages of fuzzy theory, grey prediction control, and genetic algorithm to develop a dynamical prediction control system. We pay our attention to the switch of the distinct grey prediction modes. It is different from the traditional grey prediction systems that usually focus on the selection of dynamic prediction steps. Furthermore, we also analyze the opportune moments for forecasting the system behaviors. We can know that there is not all of the control processes suitable to implement the prediction control. So we should modify the framework of the controller. Finally, we use genetic algorithms to help us for designing such a complex system. From the results of simulated experiments, we can see that the proposed methods make the performance response to track the desired trajectory well.
