UIC | Urban Transportation Center-Events: UTC Presenters at TRB 2009 Meeting

TRB 2009
88th Annual Meeting, January 11-15, Washington, DC

UTC and UTC Affiliated Presentations

Order of information:

Session number
Day, Time, Location
Session topic
Participants

More information on the TRB web site http://www.trb.org/meeting/2009/default.asp

-Session 221
Monday, January 12, 2009, 8:00 AM-9:45 AM, Hilton
Applications of Forecasting Tools and Techniques
Patrick Coleman, AECOM Consult Inc., presiding
Nazrul Islam, Federal Transit Administration, presiding

Travel Behavior and the Effects of Household Demographics and Lifestyles (09-2826)
Taha H. Rashidi, University of Illinois,Chicago
Abolfazl Mohammadian, University of Illinois,Chicago
Yongping Zhang, Wilbur Smith Associates

-Poster Session 339
Monday, January 12, 2009, 2:30 PM - 5:00 PM, Hilton
Travel Demand Forecasting Innovations
Thomas F. Rossi, Cambridge Systematics, Inc., presiding

Investigating Contextual Variability in Mode Choice Using Hierarchical Mixed Logit Model (09-2727) - B6
Liang Long, Cambridge Systematics, Inc.
Jie (Jane) Lin, University of Illinois, Chicago
Kimon Proussaloglou, Cambridge Systematics, Inc.

-Poster Session 373
Monday, January 12, 2009, 6:30 PM - 9:30 PM, Hilton
A Look at the Love Affair With the Automobile: Vehicle Ownership and Use
Xinyu Cao, University of Minnesota, Twin Cities, presiding

Competing Hazard Model of Household Vehicle Transaction Behavior with Discrete Time Intervals and Unobserved Heterogeneity (09-2832) - L9
Taha H. Rashidi, University of Illinois, Chicago
Abolfazl Mohammadian, University of Illinois, Chicago

-Poster Session 378
Monday, January 12, 2009, 6:30 PM - 9:30 PM, Hilton
Understanding Activity and Time Use Patterns
Dick Ettema, Utrecht University, Netherlands, presiding

Implementation of Scheduling Conflict Resolution Model in an Activity Scheduling System (09-2042) - L8
Joshua Auld, University of Illinois, Chicago
Abolfazl Mohammadian, University of Illinois, Chicago
Matthew J. Roorda, University of Toronto, Canada

Adapts: Agent-Based Dynamic Activity Planning and Travel Scheduling Model--A Framework (09-2037) - L7
Joshua Auld, University of Illinois, Chicago
Abolfazl Mohammadian, University of Illinois, Chicago

-Poster Session 387
Monday, January 12, 2009, 7:30 PM - 9:30 PM, Marriott
Integration, Deployment, and Maintenance of Intelligent Transportation Systems
Wei-Hua Lin, University of Arizona, presiding

Overview of Approaches to Privacy Preservation in Intelligent Transportation Systems and Vehicle Infrastructure Integration Initiative (09-0416) - C6
Caitlin D. Cottrill, University of Illinois, Chicago

-Poster Session 388
Monday, January 12, 2009, 7:30 PM - 9:30 PM, Marriott
Applications of Intelligent Transportation Systems: Monitoring and Managing Traffic

Real-Time Estimation of Urban Street Travel Time Using Buses as Speed Probes (09-2688) - A11
Wenjing Pu, Parsons Brinckerhoff, Inc.
Jie (Jane) Lin, University of Illinois, Chicago
Liang Long, Cambridge Systematics, Inc.

-Session 406
Monday, January 12, 2009, 7:30 PM - 9:30 PM, Hilton
Advances in Travel Survey Methods
Ho-Ling Hwang, Oak Ridge National Laboratory, presiding

Model-Based Synthesis of Household Travel Survey Data in Small and Mid-Size Metropolitan Areas (09-3456)
Liang Long, Cambridge Systematics, Inc.
Jie (Jane) Lin, University of Illinois, Chicago
Wenjing Pu, Parsons Brinckerhoff, Inc.

-Session 408
Monday, January 12, 2009, 7:30 PM - 9:30 PM, Hilton
Bus Systems Planning
George W. Pierlott, Mundle & Associates Inc., presiding

Planning for Bus-on-Shoulder Operations in Northeastern Illinois: Survey of Stakeholders (09-0381)
Paul Metaxatos, University of Illinois, Chicago
Piyushimita (Vonu) Thakuriah, University of Illinois, Chicago

-Session 440
Tuesday, January 13, 2009, 8:00 AM - 9:45 AM, Hilton
Applications of Advanced Computational Methods to Transportation Planning
Gary S. Spring, Merrimack College, presiding

Tour-Based Mode Choice Modeling: Using an Ensemble of Conditional and Unconditional Data Mining Classifiers (09-3281)
James P. Biagioni, University of Illinois, Chicago
Piotr M. Szczurek, University of Illinois, Chicago
Peter C. Nelson, University of Illinois, Chicago
Abolfazl Mohammadian, University of Illinois, Chicago

-Poster Session 466
Tuesday, January 13, 2009, 9:30 AM - 12:00 PM, Hilton
Advances in Alternative Fuels and Vehicles
Robert E. Larson, U.S. Environmental Protection Agency, presiding
Shannon Baxter-Clemmons, South Carolina Hydrogen and Fuel Cell Alliance, presiding

Use of Biodiesel in Railways and Its Impact on Greenhouse Gas Emissions and Land Use (09-1537) - A1
Simon Thomas McDonnell, New York University
Jie (Jane) Lin, University of Illinois, Chicago

-Poster Session 681
Wednesday, January 14, 2009, 9:30 AM - 12:00 PM, Hilton
Research Developments in School Transportation
Ann M. Dellinger, Centers for Disease Control and Prevention, presiding

School Bus Routing Problem in Large-Scale Networks: A New Approach Utilizing Tabu Search on a Case Study in Developing Countries
(09-0660)- A5
Taha H. Rashidi, University of Illinois, Chicago
Hedayat Zokaei Aashtiani, Sharif University of Technology, Iran
Abolfazl Mohammadian, University of Illinois, Chicago

-Session 770
Wednesday, January 14, 2009, 4:30 PM - 6:00 PM, Hilton
Paratransit Research: Designing Service, Analyzing Costs, and Understanding Taxi Driver Behavior
David Chia, Planners Collaborative Inc., presiding

Cost Estimation for Provision of Americans with Disabilities Act Free Special Services in Illinois (09-0380)
Paul Metaxatos, University of Illinois, Chicago
Joseph DiJohn, University of Illinois at Chicago
Lise Dirks, University of Illinois at Chicago
Karin Allen, University of Illinois at Chicago

-Session 795
Thursday, January 15, 2009, 10:15 AM - 12:00 PM, Hilton
Behavioral Considerations in Response to Energy Prices and Global Climate Change
Abolfazl Mohammadian, University of Illinois, Chicago, presiding

Session Abstracts

Travel Behavior and the Effects of Household Demographics and Lifestyles (09-2826)
Taha H. Rashidi, University of Illinois, Chicago
Abolfazl Mohammadian, University of Illinois, Chicago
Yongping Zhang, Wilbur Smith Associates

Household and individual demographics, attributes and dynamics have significant effects on their travel behavior and the overall performance of the transportation system. This study attempts to study the effects of demographic changes on the travel attributes of the household members of several homogeneous lifestyle clusters. Using the National Household Travel Survey (NHTS) 2001 data, more than twenty travel attributes including number of auto trips, trips per tour, transit usage and average commute distance are analyzed. To investigate the impact of changing demographics on household and individual level travel attributes, the best fitted distributions for a large set of travel attributes are introduced. Then the study provides a detailed comparison of the resulted distributions across different lifestyles and demographics.

In this paper a hierarchical random-coefficient mixed-logit model is applied to quantify variability in coummuters’ mode choice in the Chicago metropolitan area, especially concerning the contextual variability by the traits of census tract of residence. It is found that individual mode choice behavior varies considerably across residential locations. Moreover, the contextual effects are found to modify the marginal utilities of mode choice. Especially, in-vehicle travel time and gasoline cost are strongly affected by census tract traits (e.g., percentage of blue collar residents, ethnicity). Furthermore, random variation is present even after both contextual and individual traits are controlled for, suggesting intrinsic randomness in individual mode choice. The hierarchical structure of quantifying contextual variability proves to be a useful tool in capturing intrinsic heterogeneity in mode choice. The study findings also caution the practice of borrowing model coefficients from one region to another. Without careful examination of the borrower and lender regions, such practice may produce misleading results.

The analysis of household vehicle ownership behavior is considered as one of the most important elements of household travel behavior studies. Having a good understanding of household vehicle transaction choice can significantly increase our ability to better understand and predict household travel behavior. Applications of hazard based duration models, especially competing risk duration models, to the automobile ownership problem have drawn attention in recent years. These models can predict timing of transaction events and changes to the household vehicle fleet in a competing fashion. Despite recent technical advances in formulation of competing risk models, their applications have been very limited probably due to difficulty of obtaining longitudinal data and complexity of estimating models with multiple outcomes. Furthermore, competing hazard models of vehicle transactions typically have assumed independence across hazards. These models often overlook the effects of unobserved heterogeneity in the model. This study attempts to address these concerns by formulating and estimating a competing hazard model of vehicle transaction with multiple outcomes that also accounts for interdependencies between hazards and unobserved heterogeneity. Three transaction decisions of purchasing, trading, or disposing of a vehicle are considered and different independent duration models, competing duration models with and without unobserved heterogeneity are estimated and compared. The study accounts for the presence of heterogeneity in competing hazard models and overcomes the difficulty of its complex formulation when dealing with multiple outcomes. Furthermore, the modeling framework is flexible enough to consider transaction times that are not continuous and the failures that occur within discrete time intervals. It was shown that the new features can significantly improve the precision the model fit of a competing duration model with multiple outcomes.

This paper estimates the impact of using a set of newly developed scheduling rules which utilize a conflict resolution model on the accuracy of estimation of the resulting schedule when compared against a previous set of scheduling rules used in the TASHA model. The rule-based conflict model to be implemented was previously estimated using the CHASE© activity scheduling process data to represent the process individuals undergo when a scheduling conflict occurs within their daily activity pattern. The updated conflict resolution rules can handle a large set of conflict cases and allows for more realistic resolution types as compared to the original scheduling rules. The updated scheduling rules are then compared to the scheduling rules in TASHA. In order to compare the two, the planned portion of the CHASE data was scheduled using both systems and the results were compared to the executed portion of the activity pattern. The results show that using the modeled conflict resolution rules does give a closer fit to the actual executed activities from the survey with negligible performance impact.

This paper describes a new framework for dynamically simulating activity planning and scheduling within an activity-based microsimulation model. The dynamic activity planning framework is a process model which attempts to replicate time-dependent activity scheduling. By modeling the actual underlying activity and travel planning and scheduling processes rather than revealed activity-travel patterns, the model can represent a much wider range of travel demand management policies, especially policies which are expected to impact the planning process of individuals. In contrast with previous activity scheduling models, the proposed model considers activity scheduling steps as discrete events within the overall activity-travel simulation, and furthermore considers each attribute decision as a separate event. The paper develops a framework for modeling each activity and its attributes, and allows for a non-fixed attribute planning order, so that there is no pre-determined planning order assumed in the model. Various stages of the model that are implemented in an overall simulation framework are discussed. In addition, some initial data results from a pilot test of a new GPS-based prompted recall survey used to capture the underlying activity attribute planning process are presented and discussed in the context of the overall model framework. The initial data tend to support the hypothesis that significant variation exists in the manner in which activities are actually planned.

The goal of this paper is to explore rationales for mitigating privacy loss in intelligent transportation systems (ITS) and to provide an overview of methods proposed to accomplish this mitigation. The limitations and potential benefits of both technological and policy-oriented approaches are examined, and potential approaches to merging the two are explored in relation to the proposed Vehicle Integration Initiative (VII) program.

Utilizing transit buses as probes to detect general vehicle traffic conditions could be a real-time traffic monitoring mechanism in an urban advanced traveler information system (ATIS). The feasibility of such an application largely depends on (a) that there exist quantifiable relationships between bus traffic and general vehicle traffic and (b) that infrequent bus travel observations (constrained by the scheduled bus headway) are sufficiently sensitive to infer real-time general vehicle traffic conditions. In view that past urban bus probe studies have only focused on the first part, the present study is designed to examine the real-time sensitivity between buses and cars. A generic framework of real-time urban travel time estimation is proposed first, followed by a field study in which both light and congested traffic conditions are observed and a simulation study in which unexpected passenger demand surge is created. The generic framework is applied in both case studies and generates reasonable travel time estimates. The study findings provide insights to the feasibility of a real bus probe application in an urban traffic environment. Future studies are desired to explore bus probe performances in other unexpected traffic congestions.

Household travel data synthesis/simulation has become a very promising alternative or supplement of survey data to both small urban areas and large metropolitan regions in which data are expensive to collect or the data required to support the planning process become outdated. This paper proposes and applies model-based approaches (i.e., Small Area Estimation (SAE) methods) to synthesize household travel characteristics. The proposed methods addresse the sampling- biases concerns in the existing methods. Specifically, three SAE methods ¨C the Generalized Regression Estimators (GREG) method, the Empirical Best Linear Unbiased Predictor (EBLUP) method, and the Synthetic method ¨C an EBLUP without random area effects, are applied to synthesize household travel characteristics at both census tract and individual levels. The SAE framework of synthesizing household travel characteristics is demonstrated with the National Household Travel Survey data (NHTS) and Census Transportation Package Planning (CTPP) data in the Des Moines metropolitan area in Central Iowa. Results indicate that SAE methods are promising approaches to synthesize unbiased aggregate and disaggregate household travel characteristics by incorporating population auxiliary information and local small household travel survey data. The proposed data synthesis methods and analysis findings will provide a useful tool for practitioners, planners and policy makers in transportation analyses. The paper also points out that by linking population synthesis with the travel data simulation framework in this paper, this method could be of broad application in transportation planning.

Operating public buses on urban highway shoulders (BOS) is a strategy that is increasingly being explored as a means to fight congestion and attract more riders to public transit. Yet, little information is available regarding the institutional, legal, project development process, costs and performance of BOS operations. This paper describes the results of a study that was undertaken in northeastern Illinois in order to examine these preliminary but important issues, while keeping in mind that these issues need to be addressed to the satisfaction of a diverse group of stakeholders. This paper reports on findings from a survey of core stakeholders and documents concerns and possible resolutions that can be used, in order to engage a larger group of stakeholders at a later time in the planning phase.

This study aims to take the lessons learned from the history of applying data-mining techniques to mode choice modeling and extend it with the characteristics inherent to tour-based datasets. In doing so, a novel adaptation of existing data-mining methods is developed through the use of an ensemble of conditional and un-conditional classifiers. By defining the notion of an “anchor mode” as the mode selected on the first trip of a tour, this ensemble of classifiers is trained with and without knowledge of the anchor mode respectively. This allows the un-conditional model to make mode predictions without pre-condition for the first trip on a tour, followed by the conditional model which then makes mode predictions for the subordinate trips on a tour, given the knowledge of the selected anchor mode from the previous trip. This method was tested on the new Chicago Travel Tracker Survey dataset, and prediction performance was evaluated across four different data-mining algorithms where the best performing solution was arrived at using a combination of Naïve Bayes for the un-conditional classifier and C4.5 for the conditional classifier. Performance was measured using metrics from the field of information retrieval, and able to demonstrate an appreciable gain by using this method. For the purposes of evaluating this technique compared to traditional discrete choice methods, (un-) conditional multinomial logit models were also constructed and compared to the data-mining based solution. While the performance of the multinomial logit was reasonable, the data-mining solution proved to have better prediction performance overall.

Use of Biodiesel in Railways and Its Impact on Greenhouse Gas Emissions and Land Use (09-1537) - A1
Simon Thomas McDonnell, New York University
Jie (Jane) Lin, University of Illinois, Chicago

Policy concern about local and global pollutant emissions from Class 1 rail freight is increasing. However, long-life characteristics of locomotives hinder mitigation opportunities typically available in other sectors. This paper focuses on the potential of soybean-based biodiesel to improve rail freights environmental performance. We investigate differing rates of biodiesel penetration on total and fossil fuel energy consumption in both the well-to-pump and pump-to-wheel cycles estimating the impact on four criteria and three Greenhouse Gas (GHG) pollutants. The latter is of particular interest as the impact of this sector on global climate change has largely been overlooked. In addition, we estimate resultant potential land use impacts. We find that carbon dioxide emission reductions are modest and they come at the cost of increased levels of local pollutants. In addition, large-scale biodiesel penetration will likely result in significant land use changes. This suggests policymakers will have to look to other mitigation strategies (i.e. mandating emissions from existing rail operations or facilitating second-generation biodiesel for rail).

School Bus Routing Problem in Large-Scale Networks: A New Approach Utilizing Tabu Search on a Case Study in Developing Countries (09-0660) - A5
Taha H. Rashidi, University of Illinois, Chicago
Hedayat Zokaei Aashtiani, Sharif University of Technology, Iran
Abolfazl Mohammadian, University of Illinois, Chicago

The Vehicle Routing Problem (VRP) is one of the most complicated optimization mathematical models and in particular the School Bus Routing Problem (SBRP) is an important and practical branch of this problem. Since the number of variables and equations are vast, finding the exact solution for this problem under real conditions is difficult and only heuristic and meta-heuristic algorithms can be used to solve it. In previous studies, many heuristic and meta-heuristic algorithms were tested in order to solve VRP. However, many of them are capable of solving the problem under specific assumptions. In other words, one might find few algorithms which are capable of solving real world cases in which large-scale networks are used and limitation of bus capacity is considered. Recently, “Ejection Chain Method” (ECM) has been introduced as a heuristic algorithm which efficiently finds a new neighbor solution. In a case study in developing countries, efficiency of several heuristic algorithms including ECM along with one Meta-heuristic algorithm, Tabu Search Algorithm (TSA), is verified for solving large-scale problems. Additionally, capacity limitation, which is usually ignored in VRP and SPRP algorithms like ECM, is considered as a restricting condition in this study’s models. This study will show that neither the ECM used individually nor its combination with TSA produces feasible solutions for real-life scenarios. The authors have developed two innovative heuristic algorithms, the Construction Feasible Solutions and the Changing Algorithm that when used in combination with TSA and ECM generate a practical and efficient procedure called, the Mixed Algorithm (MA). Addressing vehicles’ capacity is mainly performed by the Construction Feasible Solutions, which works to generate feasible solutions (solutions satisfying capacity limitations as well) from the unfeasible solutions that might result from the TSA and ECM. The Changing Algorithm is responsible for generating a local optimum for every resulting feasible solution. Data from bus routing of a girls' middle school was used to show the effectiveness of the Mixed Algorithm.

Instituting a free fare for ADA complementary paratransit service in the state Illinois will expectedly increase the demand and the associated costs of providing the specialized service. This paper proposes a method to estimate such demand and costs increase. Our results show an estimated average increase in annual ADA trips between 71% and 95% in the Chicago area. The range in estimated statewide annual ADA trips increase at a 90% confidence level would be between 37% and 135%. Given previous industry free ride experiments, the latent demand exhibited by the large number of disabled persons living within ¾ mile of a fixed route and the expected diversion of wheelchair riders currently using fixed routes, we believe it is not unreasonable to expect increases in ridership approaching 100%. Compared to the (2007 dollar) baseline total statewide cost of $99.3 million, the estimated cost due to increased demand would be between $141.5 million and $202.9 million.

TRB 2009 88th Annual Meeting, January 11-15, Washington, DC

UTC and UTC Affiliated Presentations

Session Abstracts

TRB 2009
88th Annual Meeting, January 11-15, Washington, DC