臺灣博碩士論文加值系統

傳統獲得旅次分佈矩陣之方法，如家戶訪問或路邊訪問等調查成本相當昂
貴，而且有相當大的抽樣誤差。因此許多研究均朝利用路段交通量資料來
估計旅次分佈這方面著手。本論文的研究目的在於提出需要較少資料的估
計方法。本研究分兩部份討論，第一部份我們將推導均衡路網之旅次分佈
矩陣可行解空間。接著，我們將藉由改善旅次分佈模式中之參數，使估計
之旅次分佈矩陣去逼近均衡路網之旅次分佈可行解空間，以獲得最佳之參
數，並藉此估計旅次分佈矩陣。第二部份將發展 CDA模式的敏感性分析，
並將此資訊應用在設法改善 CDA模式中之參數，使估計之路段交通量去逼
進調查之部份路段交通量資料，以獲得最佳之參數，並藉此估計旅次分佈
矩陣。

Coventional methods for obtaining origin-destina- tion(OD)
trip matrix in urban areas are prone to con- siderable
sampling error. For these reasons several researchers try to
develop methods for estimating OD matrix from traffic count
data. The purpose of this thesis lies in proposing a new
estimating approach which needsless data. This reseach is
divided into two parts. The first part we derived the feasible
trip matrix set. Then, we calibrate the parameter of entropy
model by means of improving iteratively the parameter to
minimize the errors between estimated trip matrix and the
feasible trip matrix set. Finally, we estimate trip matrix. The
second part we develop a calibrating method which is based
on the sensitivity analysis of CDA model. The
calibrating method is to minimize the errors between
the estimated link flow and the observed link flow by
improving iteratively the para- meter. Finally, we estimate
trip matrix.