| Summary: | 碩士 === 逢甲大學 === 工業工程與系統管理學研究所 === 97 === Data Envelopment Analysis (DEA) is one of performance evaluation methods applying multiple inputs and multiple outputs from targeted decision making units (DMU). Traditional DEA models can only separate efficiency DMUs from non-efficiency DMUs. When one wants to completely rank DMUs or design alternatives, advanced DEA methods might be needed for this purpose. In addition to traditional DEA models, several advanced DEA models are needed for completely ranking design alternatives.
The aircraft industry is technology-intensive, experience-intensive and capital-intensive. The aircraft engine accounts for about 30% of the total aircraft product value. The aircraft engine protection system is a key device to preventing aircrafts from malfunction. When designing the engine protection systems, the designer should consider many important factors such as reliability factor, cost factor, volume factor, weigh factor, and satisfactory factor. Using many combinations of design factors, the designer wants to know how to evaluate the best one among all design alternatives by DEA models.
In this study, CCR model and BCC model are used to differentiate those efficiency and non-efficiency design alternatives. Then six advanced DEA models, including Andersen and Petersen model, Cross Efficiency model, Aggressive Cross Efficiency model, Benevolent Cross Efficiency model, Cross Reference model, and Common Weight, are analyzed and applied to completely rank those design alternatives. The study results suggest that Andersen and Petersen model and Cross Reference model give a consistent assessment result. On the other hand, Cross Efficiency model, Aggressive Cross Efficiency model and Benevolent Cross Efficiency model show a peculiar ranking result. Common Weight model can not provide a complete rank and need further investigation.
|
|---|
