Complexity analysis of traffic time series based on multifractality and complex network

Abstract

The research of traffic network becomes a hot field of traffic disciplines nowadays, to reveal the intrinsic evolution rule of traffic flow and congestion principle. In this paper, traffic flow theory and the fractal theory introduced, based on the complex network, combining with the cellular automaton, the complexity and nonlinear characteristics of traffic flow evolution are qualitatively and quantitatively analyzed. In order to study the complexity of traffic flow evolution and dynamic mechanism, with the increase of vehicle density, two methods have been used. First, multifractal detrended fluctuation analysis is used to detect multifractality of traffic time series. Second, visibility graph is used to convert time series into networks, by topology parameters to show that the traffic network is characterized with the features of scale free and small world. It is shown that this construction method is effective for traffic time series analysis.