雙限旅次分佈係指在各起點之旅次產生總量與各迄點吸引總量皆同時固定的情況下，找出起迄需求量之分佈情形。交通量指派的輸入部份須使用旅次分佈計算所得的各起迄間之運輸需求量；而從事旅次分佈工作時，又必須利用交通量指派所得的各起迄點間之旅行成本，兩者關係環環相扣，利用整合模型可完全消除兩介面間不一致性的問題。 傳統求解雙限旅次分佈與交通量指派整合問題主要有直接演算法、雙階段演算法兩種。過去研究已指出求解雙限旅次分佈與交通量指派問題雙階段演算法優於直接演算法，但兩者皆仍存在不夠精確且無法獲得合理唯一路徑解之缺點。新近發展之交通量指派演算法多半可以達到所需的精確度，但對合理路徑流量唯一解課題未見深入探討，直到Bar-Gera (2010) 提出TAPAS演算法，融入流量比例原則之概念，在達到用路人均衡的情況下，所得解處於極大熵狀態，可得到合理的路徑流量唯一解。 本研究提出新的方法，利用Chen (2011) 延伸性之概念，將雙限旅次分佈與交通量指派整合問題轉換為延伸性交通量指派問題，概念簡淺易懂，並建構極大熵雙限整合模型以及雙限超級路網。以延伸性TAPAS演算法進行範例求解與驗證，研究結果發現延伸性求解概念，不僅精確度與運算效率皆優於傳統之雙階段演算法，善用TAPAS演算法特性，能夠突破以往整合模型無法獲得合理唯一路徑流量解之瓶頸。 ;The doubly constrained trip distribution problem is to find the O-D demands assuming that both the total flow generated at each origin and the total flow attracted to each destination are fixed and known. The input data of TD(trip distribution) and TA(traffic assignment) has to depend on each other. Using the concept of combined model can eliminate the inconsistency of two interfaces. Traditionally, in order to solve the combined model, which the TD is considered along with TA, there are mainly two method: direct method and double-stage algorithm. Though the double-stage algorithm is much more accuracy and precise than the direct method in solving the combined model, they both still are not accuracy and precise enough in acquiring the unique route flow solution procedure, according to the past studies. Most of traffic assignment algorithm still couldn′t conquer multiple solutions problems, until Bar-Gera (2010) mentioned TAPAS algorithm, which add into the concept of proportionality and entropy. Here is a new method presented in the article. We utilize the concept of extend network(Chen ,2011), and regard doubly constrained combined model as extend traffic assignment problem to construct the MEUE combined model and supernetwork. Using the extend TAPAS algorithm to solve it. The results indicate that the extend TAPAS algorithm is more efficient and has a better precision than the double-stage algorithm. Besides, we can breach the bottleneck to get the unique route flow solution for combined model.
