Adaptive hybrid fuzzy rule-based system approach for modeling and predicting urban traffic flow Loukas Dimitriou a , Theodore Tsekeris a,b , Antony Stathopoulos a, * a Department of Transportation Planning and Engineering, School of Civil Engineering, National Technical University of Athens, 5 Iroon Polytechniou, 157 73 Athens, Greece b Centre for Planning and Economic Research, Athens, Greece Received 6 November 2006; received in revised form 13 November 2007; accepted 14 November 2007 Abstract This paper presents an adaptive hybrid fuzzy rule-based system (FRBS) approach for the modeling and short-term fore- casting of traffic flow in urban arterial networks. Such an approach possesses the advantage of suitably addressing data imprecision and uncertainty, and it enables the incorporation of expert’s knowledge on local traffic conditions within the model structure. The model employs univariate and multivariate data structures and uses a Genetic Algorithm for the offline and online tuning of the FRBS membership functions according to the prevailing traffic conditions. The results obtained from the online application of the proposed FRBS are found to overperform those of the offline application and conventional statistical techniques, when modeling both univariate and multivariate traffic data corresponding to a real signalized urban arterial corridor. Ó 2007 Elsevier Ltd. All rights reserved. Keywords: Traffic flow modeling; Short-term forecasting; Fuzzy rule-based systems; Global optimization; Urban networks 1. Introduction Accurate forecasting of traffic conditions in the short run, i.e. within the next few minutes, constitutes a crucial element for the efficient deployment of advanced traffic management and information systems (ATMIS) to prevent or mitigate congestion problems in metropolitan areas. Short-term traffic forecasts can be used to describe future traffic state variables, such as traffic flows, support the operation of online adaptive and transit priority signal control and route guidance systems along urban arterial streets. Several methods have been developed and implemented for the short-term prediction of traffic flow, mostly in freeways and, at a smaller extent, in urban arterial networks. 0968-090X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.trc.2007.11.003 * Corresponding author. Tel.: +30 210 7721288; fax: +30 210 7721287. E-mail addresses: lucdimit@central.ntua.gr (L. Dimitriou), fabtse@central.ntua.gr (T. Tsekeris), astath@transport.ntua.gr (A. Stathopoulos). Available online at www.sciencedirect.com Transportation Research Part C xxx (2007) xxx–xxx www.elsevier.com/locate/trc ARTICLE IN PRESS Please cite this article in press as: Dimitriou, L. et al., Adaptive hybrid fuzzy rule-based system approach ..., Transport. Res. Part C (2007), doi:10.1016/j.trc.2007.11.003