Boteler J.,MacroSys LLC
Public Roads | Year: 2011
Jennifer Boteler examines the importance of transportation libraries in assisting researchers find what they need to make sound decisions in the face of ever-expanding universe of knowledge. Generally, librarians are trained and skilled in identifying and locating accurate, relevant information. They are adept at helping researchers narrow their queries to the most essential or key documents in a field of study. At the Federal Highway Administration (FHWA) Research Library, housed at the Turner- Fairbank Highway Research Center (TFHRC) in McLean, VA, library staffers spend considerable time communicating with highway researchers in order to understand the nature of the specialized information being requested. The librarians delivered copies of rare documents in 1 to 2 days, provided access to electronic copies through library subscriptions to various full-text databases that the project otherwise could not afford to access, and found and delivered copies of unpublished documents the research team was unlikely to have found on its own.
Thakuriah P.,University of Illinois at Chicago |
Menchu S.,University of Illinois at Chicago |
Tang L.,MacroSys LLC
Transportation Research Record | Year: 2010
An individual-level mixed longitudinal data set with three separate cohorts representing three generations in U.S. society is used to examine how car ownership trends have changed among young adults (18 to 24 years of age) over a 40-year period, from the mid-1960s to the mid-2000s. A multi-level, mixed model for binary car-ownership outcomes was estimated by using maximum likelihood. After controlling for factors relating to sociodemographics, personal life, location, and labor market, individuals in the past generation considered (who were young adults in the late 1990s and early 2000s) were found to be significantly more likely to enter the car market at an earlier age. In contrast to earlier generations, by the time individuals in the last generation cleared the 18- to 24-year window, they were found to have reached car ownership levels that had been predicted by earlier studies to be the saturation levels of car ownership in a society. Minorities made greater progress in rates of car ownership between the late 1960s and the early 1980s compared with nonminorities; however, increases in the rates of car ownership slowed for minorities in the most recent generation. Young adults in urban areas led those in rural areas in the earlier generations on car ownership rates but fell behind their rural counterparts in the last generation considered. Across all three generations, vehicle ownership was lower among the unemployed compared with the employed. The differential between employed and unemployed increased with each successive generation and was almost 30% in the last generation considered.
Lu Q.,MacroSys LLC
2nd International Conference on Advanced Geographic Information Systems, Applications, and Services, GEOProcessing 2010 | Year: 2010
A single-origin multiple-destination route (SOMDR) query is to retrieve the minimum-cost routes, each of which starts from the same origin and ends at one of a set of destinations. The query is typically used for route assignment in transportation networks in GEOProcessing. In this paper, LU, a fundamental best first search algorithm and framework, is proposed to process SOMDR queries in a graph. It uses a heuristic, h-LU, estimated based on destinations yet-to-be-reached to expedite the search process following a best first way. It is a framework that can adopt different heuristics to provide optimal and sub-optimal solutions. The paper discusses how to incorporate heuristic information obtained from the problem domain into a formal mathematical theory of graph searching and how a family of search strategies can demonstrate an optimality property in a sub graph. As an example, the Euclidean distance between two vertices is used as the basis to provide a consistent heuristic, h-LU, for LU to process SOMDR queries in network distance in a transportation network and a set of experiments is performed accordingly. The result demonstrates that LU is much more efficient than Dijkstra's algorithm when the number of destinations is relatively smaller than the total number of vertices in a graph. On average, Dijkstra's algorithm expands 0.2-4.7 times more vertices and is 0.2-20.6 times slower than LU to retrieve optimal solutions when the number of destinations is up to 100. As a best first search, LU extends the capability of existing best first searches from single-destination query processing to multi-destination query processing. For SOMDR query processing, Dijkstra's algorithm is a special case of LU when no heuristic is adopted during the search process, and A* is a special case of LU when the number of destinations is one. © 2010 IEEE.
Thakuriah P.,University of Illinois at Chicago |
Tang L.,MacroSys LLC |
Vassilakis W.,University of Illinois at Chicago
4th ACM SIGSPATIAL International Workshop on Computational Transportation Science 2011, CTS'11, in Conjunction with ACM SIGSPATIAL GIS 2011 | Year: 2011
We analyze temporal and spatial variations in bus ridership that may result from using real-time bus arrival information. Using Random Effects Negative Binomial models of longitudinal average weekday ridership per bus route and controlling for operational, economic and social factors, we assess temporal variations in ridership by means of two types of time-varying coefficients: one that reflects "adjustment interval" effects after information becomes available on a route, during which users learn about information availability and potentially adapt their travel behavior, and the second, a "period" effect, that reflects changes in the underlying information and communications technology over time and ways in which people receive and use information. A k-means cluster analysis of bus stop service areas along routes that accrued the highest ridership allows us to associate the net effects of information to the sociodemographic, built environment, housing, economic, transportation and digital savviness characteristics of service areas. Four clusters of bus stops were identified: two where bus boardings gains were high after Bus Tracker, and two others where boarding gains were either modest or low. This strategy helped to determine the types of spatially-targeted Location-Based Service applications that may be developed to capitalize on basic bus arrival information. © 2011 ACM.
Tang L.,MacroSys LLC |
Ross H.,MacroSys LLC |
Han X.,MacroSys LLC
Transportation Research Record | Year: 2012
The effects of a real-time bus information system, Chicago Transit Authority (CTA) Bus Tracker, on the ridership of the CTA transit rail system in Chicago, Illinois, are examined with panel data on monthly average weekday ridership at each transit train station in the entire CTA system from January 2005 through December 2010. Other factors that might contribute to changes in CTA rail ridership (e.g., socioeconomic situation, transit service, and weather) are controlled. Results of a mixed linear model indicate that Bus Tracker service increases CTA train ridership. This increase might be attributable to the increased intermodal transfer efficiency between CTA buses and trains that results from the real-time bus information provided. Findings suggest that the complementarity effect of providing real-time bus information outweighs its substitution effect on the ridership of a connected transit rail system. Together with the findings from an earlier study (which show that Bus Tracker service helps to increase bus ridership), study results provide empirical evidence in support of the hypothesis that implementation of a real-time transit information system leads to gains in transit ridership.