大型事件的舉辦常吸引大量的旅次聚集，如何在事件發生時有效的疏散所吸引的旅次，將是一項重要的工作。為了利於疏散所吸引的旅次，在事前就必須清楚而完善地規劃設計疏散路網，讓整個事件區域的運輸路網順利進行運作。大型事件舉辦時，整體運輸路網系統設計，將影響一活動的成功與否；本研究的主要重點定位於連續型大型事件活動，針對私人運具的使用者進行疏散路網設計的理論性研究，期透過疏散路網設計模型的建立，對於連續型的大型事件活動疏散交通規劃上，提供規範與區隔大型事件活動的參與者，與非活動參與者及當地居民的道路使用權限的系統性方法，使之符合活動的疏散要求，並使非活動參與者及當地居民在交通上重大的衝擊最小。本研究屬網路設計問題，將建立雙層規劃模型，上層模型係以建立一個使總路網旅行成本最小化的系統目標為疏散路網構建策略，下層模型則考量大型事件活動參與者的疏散路徑選擇與區位選擇行為，以及非活動參與者的路徑選擇行為作為限制條件。由於本研究問題為一非線性混合0-1整數規劃問題，在求解上極為困難，在此採用模擬退火法，結合對角化法、梯度投影法等求解技術設計求解演算法，最後並以臺北市大安、中正、萬華三區主要道路所形成的測試路網，配合情境設計的數值範例，說明本模型的正確性與實用性。本研究之成果，將可提供交通管理相關權責單位，作為未來各項大型活動研擬相關疏散計畫上重要的決策參考。

Mass-event's activities usually attract a large number of trips in a short time period. Therefore, how to efficiently evacuate the activity participants which are attracted by mass-events is an important issue. When a mass-event is held, the network traffic performance highly depends on prior planning of the transportation network system. In this study, a network design model for continuous mass-events evacuation is presented. Under the network design mechanism, the traffic flow of activity participants and non-participants are distinguished. We expect in the mass-events network design model, the traffic flow of participants can be evacuated effectively, and the flow of non-participants such as local residents will be influenced and impacted as little as possible. This study deal with a network design problem. We formulate it as a bi-level programming model. The upper level problem considers how to construct an evacuation network with minimum network total travel costs. The lower level problem is one of the constraints for the upper level problem and considers the participants' evacuating route/destination choice behaviors as well as non-participants' route choice behaviors. Because the solution of this evacuation network design model is to slove non-convex and mixed 0-1 integer nonlinear programming, it is difficult to obtain. In solving the problem, the Simulated Annealing Method, Diagonalization Method and Gradient Projection Method are adopted to develop the solution algorithm for the bi-level model. The numerical test results indicate that the solution of the evacuation network design model for mass-events is accurate and practical. The result of this research is an important reference for decision-making process for government, military units and other administrations when holding mass-event related activities in the future.