Ck Wong

URBAN TRAFFIC FLOW PREDICTION USING A FUZZY-NEURAL APPROACH

Abstract This paper develops a fuzzy-neural model (FNM) to predict the traffic flows in an urban street network, which has long been considered a major element in the responsive urban traffic control systems. The FNM consists of two modules: a gate network (GN) and an expert network (EN). The GN classifies the input data into a number of clusters using a fuzzy approach, and the EN specifies the input–output relationship as in a conventional neural network approach. While the GN groups traffic patterns of similar characteristics into clusters, the EN models the specific relationship within each cluster. An online rolling training procedure is proposed to train the FNM, which enhances its predictive power through adaptive adjustments of the model coefficients in response to the real-time traffic conditions. Both simulation and real observation data are used to demonstrative the effectiveness of the method.

Lane-based optimization of signal timings for isolated junctions

This paper presents a lane-based optimization method for the integrated design of lane markings and signal settings for isolated junctions. Both traffic and pedestrian movements are considered in a unified framework. Capacity maximization and cycle length minimization problems are considered. The problems are formulated as Binary-Mix-Integer-Linear-Programs (BMILP), which are solvable by any standard branch-and-bound routine. The integer variables include the permitted movements on traffic lanes and successor functions to govern the order of signal displays, whereas the continuous variables include the assigned lane flows, common flow multiplier, cycle length, and starts and durations of green for traffic movements and lanes and pedestrian crossings. A set of constraints are set up to ensure feasibility and safety of the resulting optimal lane markings and signal settings. Numerical examples are given to demonstrate the effectiveness of the proposed method.

A parallelized genetic algorithm for the calibration of Lowry model

Abstract This paper presents a parallelized genetic algorithm for the calibration of Lowry model based on a maximum likelihood approach. A case study for the city of Hong Kong was employed for demonstrating the performance of the parallelized genetic algorithm, in terms of two commonly used performance measures: speedup and efficiency. The genetic algorithm is particularly suitable for implementation under a parallel computing environment. The parallelized version of the genetic algorithm is efficient and can be used to substantially reduce the computing time requirement for the calibration procedure. Therefore, it greatly enhances the potential applicability for large scale problems. An empirical study on the performance of the algorithm was conducted, from which an empirical formulae was developed to indicate the likely computing time in relation to the number of processors used for parallel computation.