本論文針對小型晶片自動燒錄設備，提出一個基於最佳化演算法之模糊行為決策系統來提升設備的產量，主要探討設備模擬器與行為決策系統的設計與實現。在設備模擬器的設計實現上，本論文針對小型晶片自動燒錄設備，設計實現一個模擬器來模擬設備的產量。並且依據不同的晶片燒錄時間，探討不同平台配置方式與不同燒錄器配置數量的產量變化。在行為決策系統的設計上，本論文提出一個基於最佳化演算法之模糊行為決策系統。首先，本論文針對小型晶片自動燒錄設備的特性，提出一個兩個輸入一個輸出的模糊系統架構，讓設備能夠依據目前各個燒錄器的狀態來決定下一個適當的動作。然後使用遺傳演算法(GA)與粒子群最佳化(PSO)二種最佳化演算法來自動選取一組較佳之模糊系統參數值，使得這個所對應的模糊行為策略系統可以讓設備具有較佳的產量。最後，由實驗的結果可以看出，針對不同的晶片燒錄時間，所實現的設備模擬器確實可以提供一個方法來決定一個適當的平台配置方式與一個適當的燒錄器數量。並且所設計的行為決策系統確實可以提升設備產量，並且減低異常狀況對產量所造成的影響。 In this dissertation, an optimal algorithm-based fuzzy behavior decision system is proposed for a small-sized chip automatic programming equipment to improve the production capacity of equipment. Two main designed topics of equipment simulator and behavior decision system are studied. In the equipment simulator design, a simulator is designed and implemented for the small-sized chip automatic programming equipment to simulate its production capacity. Moreover, depending on different programming time, the change of production capacity is studied for different platform configuration and different number of programmers. In the behavior decision system design, an optimal algorithm-based fuzzy behavior decision system is proposed to improve the production capacity. First of all, based on characteristics of the small-sized chip automatic programming equipment, a two-inputs-and-one-output fuzzy system structure is constructed. It allows the equipment to determine a next appropriate action based on the current status of each programmer. Then two optimal algorithms of GA and PSO are applied to automatically select a better parameter set so that the corresponding fuzzy behavior decision system can let the equipment have a better production capacity. Finally from the simulation results, we can see that the implemented simulator can indeed provide a method to decide an appropriate platform configuration and an appropriate number of programmers for different programming time. And the proposed method can indeed improve the production capacity and reduce the impact in some abnormal conditions.
