Search

There have generally been two kinds of approaches to the empirical study of human mobility. At the group level, some valuable information might be submerged in statistical noise, while due to the diversity of individual purpose and preference, there is still no general statistical regularity of human mobility at the individual level. In this paper, we considered group-level human mobility as the combination of several basic patterns and analyzed the collective mobility by category. Utilizing matrix factorization and correlation analysis, we extracted some of the stable/occasional components from the collective human mobility in the Beijing subway and found that the departure and arrival mobility patterns have different characteristics, both in time and space, under various conditions. We classified individual records into different patterns and analyzed the most likely trip distance by category. The proposed method can decompose stable/occasional mobility patterns from the collective mobility and identify passengers belonging to different patterns, helping us to better understand the origin of different mobility patterns and provide guidance for emergency management of large crowds. [ABSTRACT FROM AUTHOR]

We studied the distribution of entry time interval in Beijing subway traffic by analyzing the smart card transaction data, and then deduced the probability distribution function of entry time interval based on the Maximum Entropy Principle. Both theoretical derivation and data statistics indicated that the entry time interval obeys power-law distribution with an exponential cutoff. In addition, we pointed out the constraint conditions for the distribution form and discussed how the constraints affect the distribution function. It is speculated that for bursts and heavy tails in human dynamics, when the fitted power exponent is less than 1.0, it cannot be a pure power-law distribution, but with an exponential cutoff, which may be ignored in the previous studies. [ABSTRACT FROM AUTHOR]

Menezes and Barabasi have once developed a framework by mean–variance analysis to investigate the relationship between the average flux and fluctuations and found a scaling-law behavior in two distinct phase. By then most related researches dedicated to understanding the scaling laws through the internal/external forces of the system. However, as for subway system, the demand of transporting a large amount of passengers during peak hours is a necessity. Thus, alternating the phase or reducing the impact of external forces by cutting down the total passengers is completely unrealistic. The more operable approaches are specifically regulating the distribution of passenger flow in order to drive the system into a more stable phase. In this paper, we carried out the mean–variance analysis on subway traffic and dedicated to discover the effect of spatial–temporal distributions of passenger flow on the fluctuations in subway traffic. By analyzing the control properties of the transit networks constructed according to different passenger flow threshold , it is found that the control mode of transit networks might differ the phase change of fluctuations, which proposed a new perspective of controlling the fluctuations in subway traffic. • Transit network of metro system was constructed to character the spatial–temporal distribution of traffic. • Clear scaling-law can be observed between the mean and variance of passenger flow in the transit network. • The scale exponent vary within a day and are dominated by the traffic variation and the control mode of the transit networks. • Traffic restrictions should be given in priority to stations with large inbound and outbound traffic at morning (evening) peak hours. [ABSTRACT FROM AUTHOR]