近年來，信賴度分析在各領域日趨重要。由於信賴度分析方法能透過模擬方式瞭解系統處於不確定的狀態下，達成預期目標的機率，所以實務上許多系統（如：電力系統、給水系統與通訊網路系統等），陸續將其納為系統規劃、設計與運作之一環。在交通運輸路網中，路網信賴度分析，不僅可以作為路網績效評估的重要指標，更可進一步發現路網中的敏感性與關鍵性路段，作為日後交通管理與改善策略的重要依據。 目前國內尚無針對交通運輸網路進行信賴度分析之研究。而國外在此方面的相關研究，又多以路網連繫信賴度(Connectivity Reliability)與路網旅行時間信賴度(Travel Time Reliability)作為主要探討主軸。Chen et al.(2002)提出以路網容量信賴度作為路網績效評估的新指標，並說明此指標，較以往所提出的路網信賴度指標更具一般性。本研究參考其信賴度分析程序，並針對其進行局部修正，結合蒙地卡羅模擬、含路段流入率容量限制之動態用路人均衡模型、路網容量模型、敏感度分析與不確定分析等方法，獲得各項信賴度資訊，進一步探討路網容量信賴度。本研究之分析架構主要以蒙地卡羅模擬產生路段容量變數，獲得路網退化資料後，進行含路段流入率容量限制動態用路人均衡模型求解，並利用卓訓榮(1991)提出之廣義反矩陣法(Generalized Inverse Approach)求算敏感度分析資訊，其後結合各項信賴度分析方法，進一步獲得路網各項信賴度分析資訊進行評估。 Reliability analysis studies the probability achieving a specified objective for a certain system, which may appear in different areas such as transportation, electric power supply, water supply, and telecommunication systems. Since the performance measures of a network can be provided and critical links of a network can be identified, reliability analysis is nowadays incorporated into the framework of system planning as an essential component by many large projects. Research on reliability analysis is still in embryo stage. Though contents of reliability analyses are broad in spectrum, the prevailing papers are mainly to deal with the reliability on connectivity and travel time using the simulation related approach. Chen et al (2002) proposed the network reserve capacity as a new measure of network performance. The procedure of computing the network reserve capacity involves Monte Carlo simulation, static user equilibrium, sensitivity analysis and some statistical transformation. This procedure contains the connectivity and travel time reliability as special cases and thus has higher degree of applicability. However, the framework adopted by Chen et al (2002) is rather demanding in computation and lacks temporal consideration in the model formulation. This project proposes a new solution procedure including dynamic user equilibrium model and sensitivity analysis involving generalized inverse approach (Cho, 1991). This new approach will be elaborated in depth and validated by means of numerical example.
