No abstract

A common way to estimate dynamic origin-destination (O-D) flows is to establish and solve a bilevel optimization model. Though numerous efforts have been devoted to effectively and efficiently solving the model, challenges still exist because of the interdependence of jointly solving the upper level O-D estimation and lower level traffic assignment problems and the nonconvexity of the model. This paper presents an alternative framework for estimating dynamic O-D flows using machine learning algorithms. The framework consists of three major modules: a learner that learns the dynamic mapping patterns describing the relationship between prior O-D flows and observed link flows, an assigner that assigns a given O-D matrix to different links based on the learner, and a searcher that iteratively searches the optimal O-D solution using the assigner. A convolutional neural network is designed as the learner and trained as the assigner. Next, the algorithms to estimate a regular O-D matrix and real-time O-D flows are separately developed by using the assigner and two designed genetic algorithms built as the searcher. The framework was evaluated with a realistic network in the downtown area of Kunshan, China. The experimental studies show that the framework can achieve satisfactory estimation performances in real time. Meanwhile, it takes raw flow ranges as the prior inputs, making it robust in the case of lacking an accurate target O-D matrix. INDEX TERMS Dynamic O-D estimation, bilevel optimization, convolutional neural network, genetic algorithm.

No abstract

Identifying and detecting the travel mode and pattern of individual travelers is an important problem in transportation planning and policy making. Mobile-phone Signaling Data (MSD) have numerous advantages, including wide coverage and low acquisition cost, data stability and reliability, and strong real-time performance. However, due to their noisy and temporally irregular nature, extracting mobility information such as transport modes from these data is particularly challenging. This paper establishes a travel mode identification model based on the MSD combined with residents' travel survey data, Geographic Information System (GIS) data, and navigation data. Using the data obtained from Kunshan, China in 2017, enriched with variables on the travel mode identification, the model achieved a high accuracy of 90%. The accuracy is satisfactory for all of the transport modes other than buses. Furthermore, among the explanatory variables such as the built environment factors (e.g., the coverage rate of a bus stop) are in general more significant, in contrast with other attributes. This indicates that the land use functions are more influential on the travel mode selection as well as the level of travel demand.

Travel time reliability (TTR) is an important performance indicator for transportation systems. TTR can be generally categorized as either segment based or origin–destination (O-D) based. A primary difference between the two TTR estimations is that route information is implied in segment-based TTR estimations. Segment-based TTR estimations have been widely studied in previous research; however, O-D–based TTR estimations are used infrequently. This paper provides detailed insight into O-D–based TTR estimations and raises three new issues: ( a) How many routes do travelers usually take and what are the TTR values associated with these routes? ( b) Do statistical differences exist between route-specific and non-route-specific (NRS) TTR values? ( c) How can O-D–based TTR information be delivered? Two processes were proposed to address the issues. Three TTR measures—standard deviation, coefficient of variation, and buffer index—were calculated. The bootstrapping technique was used to measure the accuracy of the TTR measures. Approximate confidence intervals were used to investigate statistically the differences between route-specific and NRS TTR measures. A large quantity of taxicab GPS-based data provided data support for estimating O-D–based TTR measures. The results of O-D–based TTR measures showed that no statistically significant differences existed between route-specific and NRS TTR measures for most of the time periods examined. Statistically significant differences could still be found in some time periods. Travelers may take advantage of these differences to choose a more reliable route. Access to both numeric TTR values and route preference, instead of just to TTR information on segments of interest, can be beneficial to travelers in planning an entire trip.