Melih Naci AĞAOĞLU, Ergun GEDİZLİOĞLU

Link ağırlık matrislerinin belirlenmesi için çekim modeli ve logit model yaklaşımı

Link ağırlık matrisleri, ulaştırma planlamasında çok önemli bir yere sahip olan, başlangıç-son matrisi tahmin modellerinde kullanılmaktadır. Trafik sayımlarını kullanarak başlangıç-son matrisi tahmini için önerilen yaklaşımlar iki grup altında toplanmaktadır. Birinci grup ulaştırma modellerini temel almaktadır. İkinci grup ise, istatistiksel sonuç çıkarma yaklaşımlarıdır. Her iki grup için günümüze dek önerilen modellerde link ağırlık matrisleri, orantılı atama veya denge ataması kullanılarak belirlenmektedir. Şehirler arasındaki trafiği belirlemek amacıyla başlangıç-son matrisi tahmin edilmek istendiğinde, sıkışma etkisi olmadığı için orantılı atama kullanılarak link ağırlık matrisleri belirlenebilmektedir. Bu çalışmada, başlangıç-son matrisi tahmininde ihtiyaç duyulan link ağırlık matrislerinin tahmini için “çekim modeli” ve “logit model” yaklaşımı önerilmektedir.

Gravity model land logit model approach for determining link weight matrices

Link weight matrices have been used in the origin-destination estimation models. Both origin-destination matrix and traffic flows on network links are important information for transportation planning, traffic operation and control. Suggested approaches for estimation of origin-destination matrix are examined under two categories. First category is based on transportation models and second is statistical inference approaches. Link weight matrices have been determined using proportional or equilibrium assignment for suggested models in recent years. The treatment of congestion effects is an important property distinguishing various models for origin-destination matrix estimation. The models assume either that congestion can be treated exogenously (by proportional assignment) or endogenously (by equilibrium assignment). In proportional assignment case, it has been assumed that link counts and link weight matrices were independent from each other. The proportion of travelers choosing a route will not depend on congestion in the network but only on traveler and route characteristics. On the other hand, link weight matrices have been estimated under congestion effect in the equilibrium assignment. Link weight matrices have been determined using proportional assignment methods for inter-city traffic and using equilibrium assignment methods for urban traffic. In this work, gravity and logit model approach has been suggested to estimate link weight matrices for inert-city roads. Suggested models are based on proportional stochastic assignment methods and utility functions have been used instead of cost function.

PDF

___

Abrahamsson, T., (1998). Estimation of origindestination matrices using traffic counts-a literature survey, IIASA interim report, Laxenburg, Austria.
Bell, M.G.H., (1991). The estimation of origindestination matrices by constrained generalized least squares, Transportation Research, B25, 13-22.
Casey, H.S., (1955). Applications to traffic engineering of the law of retail gravitation, Traffic Quarterly, IX, 1, 23-35.
Dial, R.B., (1971). A probabilistic multi-path traffic assignment model which obviates path enumeration, Transportation Research, 5, 2, 83-11.
Erlander, S., Nguyen, S. ve Stewart, N., (1979). On the calibration of the combined distribution / assignment model, Transportation Research, B13, 259–267.
Fisk, C.S. ve Boyce, D.E., (1983). A note on trip matrix estimation from link traffic data, Transportation Research, B17, 245-250.
Jörnsten, K. ve Nguyen, S. (1979). On the estimation of trip matrix from network data, Technical Report LiTH-MAT-R-79-36, Department of Mathematics, University of Linköping, Linköping, Sweden.
Kawakami, S., Lu, H. ve Hirobata, Y., (1992). Estimation of origin-destination matrices from link counts considering the interaction of the traffic modes, Papers in Regional Science, 71, 2, 139-152.
Maher, M., (1983). Inferences on trip matrices from observations on link volumes: a Bayesian statistical approach, Transportation Research, B17, 435-447.
Ortuzar, J.D.D. ve Willumsen, L.G., (1990). Modeling Transport, 239-297, J. Wiley & Sons, England. Robillard, P., (1975). Estimating the O-D matrix from observed link volumes, Transportation Research, 9, 123-128.
Spiess, H., (1987). A maximum likelihood model for estimating origin-destination matrices, Transportation Research, B21, 395-412.
Spiess, H., (1990). A descent based approach for the OD matrix adjustment problem, Special Publication. 693, Centre de recherche sur les transports Université de Montréal, Montréal, Kanada.
Tunç, A., (2003). Trafik mühendisliği ve uygulamaları, Asil Yayın Dağıtım, Feryal Matbaacılık, Ankara
Van Zuylen, H.J. ve Willumsen, L.G., (1980). The most likely trip matrix estimated from traffic counts, Transportation Research, B14, 281-293.
Wardrop, J., (1952). Some theoretical aspect of road traffic research, Proceedings Institute Civil Engineers Part II, 325-378.
Willumsen, L.G., (1984). Estimating time-dependent trip matrices from traffic counts, In Proc. of Nineth International symposium on transportation and traffic theory, Hollanda.