| Summary: | 碩士 === 國防管理學院 === 資源管理研究所 === 87 === Cleanroom software engineering can be used to design and develop a software system with more efficient, more reliable and lower cost than others after building a usage model by using Markov chain. The model can obtain related parameter values to help a project manager to handle well the phases of R&D, design and certification simultaneously. However, there exist some disadvantages of current prediction methods in usage distribution. Hence, the techniques of Rank Weight, Analytic Hierarchy Process and Rating will be proposed under incomplete information by Decision Theory in this study to overcome above disadvantages. Their results also will be identified by statistical approaches at the same time. Moreover, the technique of integrated statistical test that can analyze large systems which decomposed with function modules, will be proposed under the testing process by using Mokov chain analysis. The proposed methods can calculate the system reliability of the complex large systems properly and rapidly. One illustrate example will be discussed detailed in this thesis. This kind of approach can be implemented in national defense technology and information industry fields to improve the quality and reliability of software products.
|
|---|
