陳中平Chen, Chung-Ping臺灣大學:電子工程學研究所劉繼蔚Liu, Chi-WeiChi-WeiLiu2010-07-142018-07-102010-07-142018-07-102008U0001-2008200810051300http://ntur.lib.ntu.edu.tw//handle/246246/189110近年的研究趨勢中，統計分析已成為一廣泛受到注意與重視之主題。在本作品中, 藉由靜態時序分析此一應用，我們提供了另一種基於數學推論，可據以進行統計分析的觀點。以實驗模型為基本,利用統計靜態時序分析中的積分法，我們所提出的方法在其實驗模型滿足以下假設時可證明其正確性：(1)其模型滿足數學上之well-defined的性質；(2)其實驗模型所定義之自變數為相互獨立；(3)所定義之自變數可分成二組無交集且無遺漏之分割，令之為 A1 和 A2，並且存在一映成函數，其定義域為待測之統計特性與A2 之聯集，而其值域為A1。Statistical Analysis draws much research attention in recent years. In this work, with the static timing analysis as target application, a mathematical analysis is made to provide another viewpoint of its statistical result. Starting from the experiment model, a statistical analysis approach based on the integration method is provided and proven to be exact with respect to the model under these requirements for the model: (1) the model is well-defined; (2) the model is based on mutually independent variables; (3) there is at least a bi-partition of independent variables, says A1 and A2, such that there’s an onto function from the union of A2 and properties to A1.Chapter 1 Introduction 1hapter 2 Preliminary and Related Work 3.1. Static timing analysis (STA) 3.2. Statistical static timing analysis (SSTA) 5.2.1. Block-based SSTA 8.2.2. Path-based SSTA 9.3. Slope Propagation 10.4. Monte Carlo method and SSTA 10hapter 3 Proposed Method 12.1. Model, Response, and general overview 13.2. Models and Cascading of Functions 16.3. Models and Particular Property 18.4. From deterministic model to statistical result 20.5. Examples and Simulation Results 27hapter 4 Discussion 38hapter 5 Conclusion 44525235 bytesapplication/pdfen-US統計靜態時序分析多變量統計分析考慮信號轉換時間SSTAMultivariate statistical analysisSlew-awarethesishttp://ntur.lib.ntu.edu.tw/bitstream/246246/189110/1/ntu-97-R95943153-1.pdf