鐵路運輸的列車服務可靠度不僅是影響服務品質重要的因素之一，更是備受社會大眾所期望。列車延滯時間是評估列車服務可靠度的主要指標之一，過去研究中認為列車流量與列車平均延滯有密切之關係，但大多是探討列車排班的延滯而非實際營運的延滯；另外，國內對於相關議題的研究較為缺乏，也欠缺實務驗證，因此本研究分別發展了「迴歸模式」和「模擬模式」探討兩者的關係，經由臺鐵實際營運資料進行分析與驗證後發現，列車延滯隨著流量的增加成非線性的成長，而且「模擬模式」明顯具有較佳的解釋能力。

Train reliability is an important factor that affects service quality, and passengers usually expect reliable service. Delay is one of the major indicators to evaluate train reliability. Previous studies indicated that there exists a close relationship between train traffic flow and average delay. However, previous studies usually focused on train delay in scheduling based on a feasible timetable rather than delay in real operation. In addition, research on train delay and traffic flow in Taiwan is rare and lacking practical analysis. Thus, this study develops several regression models and a simulation model to analyze their relation. Through the analysis of the real data collected from Taiwan Railway Administration and model validation, we find that train delay increases nonlinearly as traffic flow grows, and that the simulation model has better explanatory capabilities than the regression models.