Time-dependent partitioning of urban traffic network into homogeneous regions

Congestion in urban areas constitutes an important problem that affects people in explicit but also implicit ways. Current research literature on Urban Traffic Estimation has shown that homogeneous distribution of vehicle density along the links of urban traffic networks plays an important role in the derivation or even the existence of the so-called Urban-Scale Macroscopic Fundamental Diagram or MFD in short. This Urban-Scale MFD can provide information that facilitates the application of perimeter traffic control strategies. In this paper, we implement a partitioning of an urban road network into homogeneous regions based on historical traffic information. Using prior information, we make informed decisions about the selection of the region on which the urban road network is based on, as well as the particular time periods for which the partitioning is to be implemented. We make use of weighted k-means, k-harmonic means and normalized spectral clustering techniques to successfully partition the region into clusters defined by low link density variability, while ensuring that the resulting partitions are spatially cohesive.

[1]  Nikolas Geroliminis,et al.  Macroscopic modeling of traffic in cities , 2007 .

[2]  Nikolaos Geroliminis,et al.  Properties of a well-defined Macroscopic Fundamental Diagram for urban traffic , 2011 .

[3]  C. Daganzo,et al.  Macroscopic relations of urban traffic variables: Bifurcations, multivaluedness and instability , 2011 .

[4]  Siamak Ardekani,et al.  Characterizing Traffic Conditions in Urban Areas , 1984, Transp. Sci..

[5]  Michael I. Jordan,et al.  On Spectral Clustering: Analysis and an algorithm , 2001, NIPS.

[6]  N. Geroliminis,et al.  Existence of urban-scale macroscopic fundamental diagrams: Some experimental findings - eScholarship , 2007 .

[7]  Carlos F. Daganzo,et al.  Urban Gridlock: Macroscopic Modeling and Mitigation Approaches , 2007 .

[8]  Dirk Helbing,et al.  The spatial variability of vehicle densities as determinant of urban network capacity , 2009, Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences.

[9]  Ulrike von Luxburg,et al.  A tutorial on spectral clustering , 2007, Stat. Comput..

[10]  Sejoon Lim,et al.  City-scale traffic estimation from a roving sensor network , 2012, SenSys '12.

[11]  Toru Nakamura.  WHITE PAPER, European transport policy for 2010 : time to decide , 2004 .

[12]  Ali S. Hadi,et al.  Finding Groups in Data: An Introduction to Chster Analysis , 1991 .

[13]  I. Prigogine,et al.  A Two-Fluid Approach to Town Traffic , 1979, Science.

[14]  Umeshwar Dayal,et al.  K-Harmonic Means - A Data Clustering Algorithm , 1999 .

[15]  Greg Hamerly,et al.  Alternatives to the k-means algorithm that find better clusterings , 2002, CIKM '02.

[16]  P. Rousseeuw.  Silhouettes: a graphical aid to the interpretation and validation of cluster analysis , 1987 .

[17]  Nikolaos Geroliminis,et al.  On the spatial partitioning of urban transportation networks , 2012 .