基于神经网络的OD分布矩阵反推方法

摘要/Abstract

摘要：

以道路网络的路段流量为基础进行OD分布矩阵估计.与以往文献不同的是本文应用了多层前馈神经网络的方法.由于路段流量与相关的OD矩阵分布之间存在连续性关系，这为神经网络模型的逼近特性提供了可行性.本文的方法适用于OD分布矩阵的实时校正.在已知OD分布矩阵的前提下，对两种情境——试验网络和实际Naples农村道路网进行仿真分析.主成分分析法的应用减少了变量个数并有利于改进输入数据.估计误差相对较低，与分析方法相反的是处理的时间几乎是实时的，因此这种方法可用于动态交通管理.本文的神经网络方法在误差和计算时间方面优于传统商业软件得到的OD估计结果.

关键词: 城市交通, OD分布矩阵估计；神经网络；主成分分析法；路段流量；方差稳定性

Abstract:

This paper attempts to deal with traffic Origin Destination (OD) matrix estimation starting from the measurements of flow on road network links. It proposes a different approach from published articles to date, by applying multilayer feedforward neural networks. Since the relationship between link flow and the related OD matrix is continuous, it is possible to use the well known approximation property of Neural Network models. The method is proposed for a realtime correction of the OD matrix. Two application scenarios were developed: a trial network and an actual rural network were both simulated by a microsimulator that assigns known OD matrices. A Principal Component Analysis (PCA) technique was used to reduce the amount of variables and to achieve improved significance for input data. The estimated error was relatively low and, as opposed to analytical approaches, the processing time was almost in real time, making this approach suitable for applications in dynamic traffic management. Comparisons with results obtained by an OD estimation commercial program show better performance in the NN approach both as regards error and computing time.

Key words: urban traffic, OD matrix estimation, neural networks, PCA(Principal Component Analysis), link flow, variance stabilization

中图分类号:

U491
U268.6

MUSSONE Lorenzo, MATTEUCCI Matteo. 基于神经网络的OD分布矩阵反推方法[J]. 交通运输系统工程与信息, 2013, 13(4): 84-93.

MUSSONE Lorenzo, MATTEUCCI Matteo. OD Matrices Network Estimation from Link Counts by Neural Networks[J]. Journal of Transportation Systems Engineering and Information Technology, 2013, 13(4): 84-93.

导出引用管理器 EndNote|Ris|BibTeX

链接本文: http://www.tseit.org.cn/CN/abstract/abstract18617.shtml

http://www.tseit.org.cn/CN/Y2013/V13/I4/84

参考文献

[1]Bittanti S. Time series and linear systems[M]. Berlin: SpringerVerlag, 1986.

[2]Bishop C M. Neural networks for pattern recognition[M]. Oxford, New York: Oxford University Press, XVII, 1995：482. 

[3]Mussone L. A review of feedforward neural networks in transportation research[M]. E&I, 6/1999, SpringerVerlag, 1999：360-365.

[4]Gong Z. Estimating the urban OD matrix: A neural network approach[J]. European Journal of Operational Research 106, 1998:108-115.

[5]Van Zuylen H J, Willumsen L G. The most likely trip matrix estimated from traffic counts[J]. Transpn. Res. B, 1980, 14: 281-293.

[6]Fisk C S. On combined maximum entropy trip matrix estimation with user optimal assignment[J]. Transpn. Res. B, 1988, 22: 66-79.

[7]Florian M, Chen Y. A coordinate descent method for the bilevel OD matrix adjustment problem[J].Int. Trans. Opl. Res. 2（2）:165-179.

[8]Cascetta E. Estimation of trip matrices from traffic counts and survey data: a generalized least squares estimators[J]. Transpn. Res, B, 1984,18: 289-299. 

[9]Bell M G H. The estimation origindestination matrices by constrained generalized least squares[J].Transpn. Res. B, 1991, 25: 115-125. 

[10] Cascetta E, Nguyen S. A unified framework for estimating or updating origin/destination matrices from traffic counts[J]. Transpn. Res. B, 1988, 22: 437-455.

[11]Van der Zijpp N. Dynamic origindestination matrix estimation on motorway networks[D]. Department of Civil Engineering, Delft University of Technology, Delft, The Netherlands, 1996. 

[12]Chang G L, Wu J. Recursive estimation of timevarying OD flows from traffic counts in freeway corridors[J]. Transportation Res. B, 1994, 28(8): 141-160.

[13]Okutani I. The Kalman filtering approaches in some transportation and traffic problems[C]//N. H. Gartner, N. H. M. Wilson (Eds.), Transportation and Traffic Theory. Elsevier, New Yorkand Proc. Tenth Int. Symposium Transp. and Traffic Theory. July 810. MIT, Cambridge, MA, 1987:397-416.

[14]Ashok K, BenAkiva M E. Dynamic origindestination matrix estimation and prediction for realtime traffic management systems[C]//Daganzo, C.F. (Ed.), Transportation and Traffic Theory. Elsevier, Amsterdam, 1993.

[15]Bierlaire M, Crittin F. An efficient algorithm for realtime estimation and prediction of dynamic od tables[J]. Operations Research, 2004, 52(1): 116-127.

[16]Ashok K, BenAkiva M. Estimation and prediction of timedependent origindestination flows with stochastic mapping to path flows and link flows[J]. Transportation Sci, 2002, 36(2): 184-198.

[17]Chang G L, Tao X. An integrated model for estimating timevarying network origindestination distributions[J]. Transportation Res, A 1999, 33: 381-399.

[18]Van Aerde M, Rakha H， Paramahamsan H. Estimation of origindestination matrices: Relationship between practical and theoretical considerations (conference paper )[J]. Transportation Research Record, 1831, 2003:122-130.

[19]Yang H, Akiyama T, Sasaki T. A neural network approach to the identification of real time origindestination from traffic counts[J]. In Proceedings of the International Conference on Artificial Intelligence Application in Transportation Engineering, 1992:253-269. 

[20]Yang H, Akiyama T, Sasaki T. Estimation of timevarying origindestination flows from traffic counts: A neural network approach[J]. Mathematical and Computer Modeling, 1998, 27(9-11):323-334.

[21]Kikuchi S, Tanaka M. Estimating an origindestination table under repeated counts of inout volumes at highway ramps: use of artificial neural networks[J]. Transportation Research Record, 1739, 2000:59-66.

[22]Bifulco G N.Road transport systems in information society (in Italian) [M]. Aracne: Roma, Italy, 2004.

[23]Montgomery D C, Peck E A. Introduction to linear regression analysis[M]. New York, NY: John Wiley & Sons, 1992.

[24]Scott D W. Multivariate density estimation[M]. New York: Wiley, 1992. 

[25]Lebart L, Morineau A, Tabard N. Techniques de la description statistique: méthodes et logiciels pour l’analyse des grands tableaux[M]. Paris: Dunod, 1977.

[26]Matteucci M, Spadoni D. Evolutionary learning of rich neural networks in the Bayesian model selection framework[J]. Int. J. Appl. Math. Comput. Sci., 2004, 14(3): 423-440.