Transportation Research and Education Center

The video begins at 0:16.

Abstract: This seminar will introduce travel models to non-modelers. It will build off the previous seminar, which introduced models in general, and discuss two primary approaches to travel modeling – four-step aggregate models and activity-based disaggregate models. The inputs, basic model methodology, and outputs of each approach will be discussed. An example of each approach will be discussed as well. The goal of the seminar is to introduce key concepts, basic differences between the two approaches, and discuss the benefits and shortcomings of each approach, with a focus on application.

Speaker Bio: Ben Stabler is a supervising planner with Parsons Brinckerhoff who specializes in planning modeling systems development. Ben has worked locally, as well as internationally, on numerous four-step and activity-based travel demand and land use modeling systems and has presented at various conferences, including TRB, the TRB Planning Applications Conference, and the Innovations in Travel Modeling conference. He is a certified GIS Professional and has worked in travel forecasting for Oregon DOT as well as PTV – the makers of VISUM and VISSIM. Ben is a member of the TRB Urban Transportation Data and Information Systems Committee (ABJ30) and is an active member of the Oregon Modeling Users Group.

Watch video

View slides

Summary: A growing concern related to large-truck crashes has increased in the State of Texas in recent years due to the potential economic impacts and level of injury severity that can be sustained. Yet, studies on large truck involved crashes highlighting the contributing factors leading to injury severity have not been conducted in detail in the State of Texas especially for its interstate system.  In this study, we analyze the contributing factors related to injury severity by utilizing Texas crash data based on a discrete outcome based model which accounts for possible unobserved heterogeneity related to human, vehicle and road-environment. We estimate a random parameter logit model (i.e., mixed logit) to predict the likelihood of five standard injury severity scales commonly used in Crash Records Information System (CRIS) in Texas – fatal, incapacitating, non-incapacitating, possible, and no injury (property damage only). Estimation findings indicate that the level of injury severity outcomes is highly influenced by a number of complex interactions between factors and the effects of the some factors can vary across observations. The contributing factors include drivers’ demographics, traffic flow condition, roadway geometrics, land use and temporal...

The video begins at 3:30.

Abstract: We all use abstractions of reality to help understand the world around us, synthesize knowledge, and to predict the consequences of our actions. These range from ad hoc mental models to highly complex mathematical creations. In this discussion we'll examine the motivations for building formal models, with particular focus on the types of models that will be explored in this course. Several different modeling approaches will be compared, along with the strengths and limitations of each. Some important questions that builders and consumers of models should ask will be covered, as well as ideas for building more useful and informative models. A discussion on how to judge the validity of a model will round out the discussion.

Speaker Bio: Rick Donnelly has over 25 years of experience in the modeling and simulation of transportation systems, from the urban to national level. His current interests include agent-based modeling of freight and logistics, integrated land use-transportation models, and dynamic network modeling. Rick leads the travel modeling and simulation practice at Parsons Brinckerhoff, an international civil and transportation engineering consultancy. He is also a senior fellow at the University of Melbourne, where he earned his doctorate in engineering, and a visiting scholar at Los Alamos National Laboratory.