Zhengyu Duan

Bike-Sharing-A new public transportation mode: State of the practice & prospects

Bike-Sharing, as a new green public transportation mode, has been developed in several western cities. In this paper we systematically describe the development process and the state of art for the Bike-Sharing Program in the world. We investigated the characteristics of bicycle, business model, operating mechanism, current problems for Bike-sharing programs of the three metropolitan cities in Europe, and we also discussed the development mode and applicability of Bike- Sharing program. Then we provide guidelines thats where and how to establish the Bike-Sharing program in developing country with the consideration of the weather and topography of the target city. Finally, we discuss how to integrate the bike-sharing to existing transport mode and the future for bike-sharing program.

How to Find the Optimal Paths in Stochastic Time-dependent Transportation Networks?

This paper conducted a theoretical study on finding optimal paths in stochastic time-dependent (STD) transportation networks. A stochastic consistent network with STD link travel times was built. The methodology of robust optimization was applied to evaluate the paths for a priori optimization without requiring the probability distribution of travel times. The paths with greatest robustness, namely minimum upper bounds of travel times were defined as the optimal path. Under the stochastic consistent condition, the STD robust optimal path model can be converted into solving a time-dependent shortest path problem in a FIFO network. Then extended Dijkstras algorithm can be applied to solve the simplified STD problem with computation complexity O(n2). In the field experiment, several tests were conducted on finding robust optimal paths in a sampled STD transportation network of Shanghai, China. The numerical results confirmed the validity of the proposed approach and verified that the natural extension of conventional Dijkstras algorithm can solve the STD robust optimal path problem as efficiently as a static network problem.

TRAFFIC CONGESTION ANALYSIS OF SHANGHAI ROAD NETWORK BASED ON FLOATING CAR DATA

With the rapid growth of urban traffic, the contradiction between traffic demand and supply has become increasingly conspicuous, and traffic congestion has become a normal state. How to identify the frequent congested road sections, estimate their influence to the entire road network; and how to improve the connectivity and accessibility of the whole road network through local traffic reformation, have become important issues to transportation planners and managers. In this paper, the authors take Shanghai as an example. The authors analyze the spatio-temporal characteristics of urban traffic congestion, and identify the frequent congested sections in the network, using floating car data. Furthermore, the authors revealed the correlation of spatial distribution and the spatial concentration characteristic of congestion, by global and local spatial autocorrelation analysis.

Urban Freight Management with Stochastic Time-Dependent Travel Times and Application to Large-Scale Transportation Networks

This paper addressed the vehicle routing problem (VRP) in large-scale urban transportation networks with stochastic time-dependent (STD) travel times. The subproblem which is how to find the optimal path connecting any pair of customer nodes in a STD network was solved through a robust approach without requiring the probability distributions of link travel times. Based on that, the proposed STD-VRP model can be converted into solving a normal time-dependent VRP (TD-VRP), and algorithms for such TD-VRPs can also be introduced to obtain the solution. Numerical experiments were conducted to address STD-VRPTW of practical sizes on a real world urban network, demonstrated here on the road network of Shenzhen, China. The stochastic time-dependent link travel times of the network were calibrated by historical floating car data. A route construction algorithm was applied to solve the STD problem in 4 delivery scenarios efficiently. The computational results showed that the proposed STD-VRPTW model can improve the level of customer service by satisfying the time-window constraint under any circumstances. The improvement can be very significant especially for large-scale network delivery tasks with no more increase in cost and environmental impacts.

Stochastic time-dependent vehicle routing problem: Mathematical models and ant colony algorithm:

This article addresses the stochastic time-dependent vehicle routing problem. Two mathematical models named robust optimal schedule time model and minimum expected schedule time model are proposed for stochastic time-dependent vehicle routing problem, which can guarantee delivery within the time windows of customers. The robust optimal schedule time model only requires the variation range of link travel time, which can be conveniently derived from historical traffic data. In addition, the robust optimal schedule time model based on robust optimization method can be converted into a time-dependent vehicle routing problem. Moreover, an ant colony optimization algorithm is designed to solve stochastic time-dependent vehicle routing problem. As the improvements in initial solution and transition probability, ant colony optimization algorithm has a good performance in convergence. Through computational instances and Monte Carlo simulation tests, robust optimal schedule time model is proved to be better than minimum expected schedule time model in computational efficiency and coping with the travel time fluctuations. Therefore, robust optimal schedule time model is applicable in real road network.

A framework for spatial interaction analysis based on large-scale mobile phone data

The overall understanding of spatial interaction and the exact knowledge of its dynamic evolution are required in the urban planning and transportation planning. This study aimed to analyze the spatial interaction based on the large-scale mobile phone data. The newly arisen mass dataset required a new methodology which was compatible with its peculiar characteristics. A three-stage framework was proposed in this paper, including data preprocessing, critical activity identification, and spatial interaction measurement. The proposed framework introduced the frequent pattern mining and measured the spatial interaction by the obtained association. A case study of three communities in Shanghai was carried out as verification of proposed method and demonstration of its practical application. The spatial interaction patterns and the representative features proved the rationality of the proposed framework.

Using Longitudinal Mobile Phone Data to Understand the Stability of Individual Travel Patterns

Most travel demand models assume that individuals’ daily travel patterns are stable or follow a fixed routine. This hypothesis is being questioned by more and more researchers. In this study, longitudinal mobile phone data were used to study the stability of individual daily travel patterns from three aspects, including activity space, activity points, and daily trip-chain patterns. The activity space was represented by the number of nonhome activity points, the radius of nonhome activity points, and the distance from home. The visitation pattern of activity points was analyzed by entropy and predictability measures. The stability of trip-chain patterns was described by the number of distinct trip chains, the typical trip chain, and the typical trip-chain ratio. Analysis of 21 days of mobile phone data from three communities in Shanghai, China, revealed that individuals’ daily travel patterns showed considerable variation. Although individuals’ visitation patterns to activity points were very regular, the day-to-day variations of individual trip-chain patterns were quite significant. On average, an individual exhibited about eight types of daily trip chains during the 21-day period. The daily travel patterns of residents in the outskirts were more stable than those of residents in the city center. Individuals’ travel patterns on weekdays were more complex than those on weekends. As individuals’ activity spaces increased, the stability of their travel patterns decreased.

An Approach to Analyze Human Activity Patterns Based on Cellular Phone Data: A Case Study of Jinhe New Town in Shanghai

In the urbanization process of big cities in China, the rise of satellite towns, the migration of manufacturing and the relocation of residents to the suburban areas have accelerated the separation between workplace and residence and brought enormous changes to the activity patterns of Chinese cities. In this paper, the authors propose a novel and data-driven method of extracting individuals’ daily activities and identifying “anchor points” (home and workplace) from mobile phone data and survey data, and apply it to the Jinhe new town. People in study area are classified into three groups, and use time-geographic concept to depict individual activity pattern of each group. Furthermore, the authors focus on the residents with obviously separated home and workplace caused by suburbanization. For representing the cluster of these people’s activities in space-time, kernel density estimation is used to detect the intensity in space-time. The activity density profile facilitates finding the spatio-temporal characteristic of the demand deriving from suburban residents. The study shows that mobile phone data allows analyzing human activity pattern in space-time at very detailed scale but also require other data resources for comprehensiveness and visualization of all people across the city.

Vehicle Routing Planning in Dynamic Transportation Network Based on Floating Car Data

The traditional research on vehicle routing planning was mostly on the assumption that link travel time is constant, but the traffic conditions in real road network are often fluctuant. In order to meet the requirements of fast and efficient delivery, it is necessary to study vehicle routing planning in dynamic transportation network. In recent years, time dependent vehicle routing problem (TDVRP) which considers traffic conditions attracted more and more scholars attention. However, most studies on TDVRP are based on simple test network, and assumed all vehicles depart from the depot at a fix time. In this paper, we studied TDVRP based on floating car data. We gave a mathematical model for TDVRP, and represented the dynamic network as a first in first out (FIFO) network by time dependent function of travel speed. Then, we designed a routing construction algorithm named DTO-NNC algorithm for TDVRP. Moreover, we constructed a test instance of 100 customers based on floating car data in the road network of Shanghai, and solved it in the case of fixed departure time and variable departure time. Through the instance, DTO-NNC algorithm has been proven efficient in real road network.