Behavioral Models for Land Use, Mobility and Energy and Resource Use

Project Manager(s): 

Milan Lovric (milan.lovric@smart.mit.edu)
Sebastian Raveau (sebastian.raveau@smart.mit.edu)

Sponsor: 

Singapore National Research Foundation

Team: 

Moshe Ben-Akiva, Maya Abou-Zeid, Song Gao, Francisco C. Pereira, Milan Lovric, Sebastian Raveau, Carlos Carrion, Varun Pattabhiraman, Li Siyu, Tan Rui, Jing Ding.

Start Date: 

January 2011

Research Highlights: 

  1. Implementation of Activity-based models using the Day Schedule approach with Household Survey Data from Singapore.
  2. Extending the Activity-based travel demand models using the Day Schedule approach by including well-being measures as indicators of utility.
  3. Development of Need-Based Utility Maximizing models which constitute an alternative approach to activity pattern generation models.
  4. Estimation of travel time models based on GPS and Household survey data, and the application of these models in time of day choice modeling.
  5. Two stage choice model for daily activity patterns with grammar based representation and customized choice sets.
  6. Development of Adaptive Route Choice model including routing policy choice set generation using taxi and smart phone based GPS data.
  7. Development of Public transport route choice model with access and egress modes using smart card and other data sources.

Abstract: 

Understanding and modeling behavior is essential for urban planning and project evaluation. It is necessary to identify, quantify and model all the relevant factors influencing individual behavior and decision-making processes. These decisions can be on a long term scale (e.g. household and firms location), a medium term scale (e.g. day-to-day activity patterns) or a short term scale (e.g. travel alternatives for particular trips). This project seeks to develop state-of-the-art behavioral models in the context of agent-based simulation.

Description: 

To plan sustainable future urban mobility systems, we need a set of forecasting tools to help make well-informed and consistent assessments of future conditions under various scenarios. Behavioral models are at the heart of the approach as these models are generally at the core of agent-based simulators. The objective is to develop state-of-the-art models to understand and forecast different behavioral rationales of households and firms.

The behavioral models of Future Urban Mobility strive to realistically model agents' decision-making at different levels (short-, medium-, and long-term).  Short-term models deal with the immediate decisions of drivers (e.g. acceleration, lane changing, and gap acceptance), pedestrians (e.g. acceleration, and direction of walking), and other travelers. Mid-term models deal with planning day-to-day decisions. These decisions are typically driven by the activities in which the travelers partake. Long-term models encompass the market interactions of firms, developers, and consumers as they build, acquire, and sell property, and also other decisions such as labor participation and location of employees and employers.

Behavioral models are complex and difficult to develop since their frameworks must be (1) flexible to capture a variety of agents' decision-making processes and (2) tractable in order to adapt to available data. In some cases, the behavioral modelers may also need to support the models by designing instruments (e.g. surveys) to collect the necessary data.

Currently, we are addressing several challenges including: the development of more behaviorally realistic models at the medium-term level that consider the frequency of activity participation (e.g. travelers do shopping three times a week); the integration of travelers’ adaptive behavior (i.e. travelers adjust their routing decisions based on current and/or future information) to route choice modeling; and models that capture the behavior of developers, firms, and consumers as they interact in the housing market.

 

References: 

  1. Tan, R., Adnan, M., Lee, D.H. and Ben-Akiva, M. (2015). A New Path Size Formulation in Path Size Logit for Route Choice Modeling in Public Transport Networks. 94th Annual Meeting of the Transportation Research Board, Washington, D.C., U.S.A. January 2015. [forthcoming in Transportation Research Record Journal]
  2. Siyu Li, Francisco Camara Pereira, Der-Horng Lee, Moshe E. Ben-Akiva (2015) A Two-Stage Choice Model for Daily Activity Patterns with Dynamic Choice Sets. In proceedings of the International Choice Modeling Conference 2015, Austin, Texas
  3. Ding, J., Gao, S., Jenelius, E., Rahmani, M., Pereira, F., and Ben-Akiva, (2015) M. Modeling Adaptive Route Choice in Stochastic Networks. The 14th International Conference on Travel Behaviour Research (IATBR), London, UK (Accepted).
  4. Ding, J., Gao, S., Jenelius, E., Rahmani, M., Pereira, F., and Ben-Akiva, M. A (2015), Latent-Class Routing Policy Choice Model with Revealed Preference Data. The 94th Annual Meeting of the Transportation Research Board, Washington, DC.
  5. Pattabhiraman, V., Carrion, C., Raveau, S., Abou-Zeid, M. and Ben-Akiva, M. (2015). A Needs-based Model of Activity Duration, Location Choice, and Frequency. Transportation Science (under review).
  6. Ding, J., Gao, S., Jenelius, E., Rahmani, M., Huang, H., Ma, L., Pereira, F., and Ben-Akiva, M.(2014), Adaptive Route Choice Models: Specification, Choice Set Generation, and Estimation - Case Study in Stockholm, Sweden. 3rd INFORMS Transportation Science and Logistics Society Workshop, Chicago IL.
  7. Ding, J., Gao, S., Jenelius, E., Rahmani, M., Huang, H., Ma, L., Pereira, F., and Ben-Akiva, M. (2014), Routing Policy Choice Set Generation in Stochastic Time-Dependent Networks: Case Studies for Stockholm and Singapore. The 93rd Annual Meeting of Transportation Research Board, Washington, DC.
  8. Ding, J., Gao, S., Jenelius, E., Rahmani, M., Huang, H., Ma, L., Pereira, F., and Ben-Akiva, M. (2013), Routing Policy Choice Set Generation in Stochastic Time-Dependent Networks: Case Studies for Stockholm and Singapore. The Annual Meeting of the Institute for Operations Research and Management Science, Minneapolis, MN.
  9. Pattabhiraman, V., Ben-Akiva, M., and Abou-Zeid, M. (2013) “A Needs-Based Model of Activity Location, Duration, and Frequency”, paper presented at the 13th World Conference on Transportation Research, Rio de Janeiro, Brazil.
  10. Siyu Li, Der-Horng Lee (2013) Learning Daily Activity Pattern with Probabilistic Grammar. In Proceedings of the 93rd Annual Meeting of the Transportation Research Board, Washington, D.C.
  11. Li, S., Carrion, C., Abou-Zeid, M. and Ben-Akiva, M. (2013). Activity-based Travel Demand Models for Singapore: Application and Innovations. 18th International Conference of Hong Kong Society for Transportation Studies, Hong Kong, China. December, 2013.
  12. Li, S., Enam, A., Abou-Zeid, M., and Ben-Akiva, M. (2013). Travel Time Modelling with GPS and Household Survey Data. 92nd Annual Meeting of the Transportation Research Board, Washington, D.C., U.S.A. January 2013
  13. Pattabhiraman, V., Ben-Akiva, M., and Abou-Zeid, M. (2012) “A Needs-Based Utility Maximizing Model of Activity Location, Duration, and Frequency”, paper presented at the 91st Annual Meeting of the Transportation Research Board, Washington, DC.
  14. Pattabhiraman, V., (2012), A Needs-based approach for activity generation for travel demand analysis, Masters Thesis, Submitted to Department of Civil and Environmental Engineering, Massachusetts Institute of Technology. 

The SimMobility and Behavioral Models projects are related. To read more information about SimMobility, please click SimMobility.