Agent-based Dynamic Activity Planning and Travel Scheduling Model: Data Collection and Model Development.

摘要

The research detailed in this thesis focuses on the creation of the Agent-based Dynamic Activity Planning and Travel Scheduling (ADAPTS) travel demand model for use in policy analysis, and on the collection of new activity planning data to aid the development of the model. The ADAPTS model is an agent-based microsimulation model which estimates travel behavior for each individual in the model region and which is combined with a disaggregate travel simulator. As such the model is both a microsimulation of travel demand as well as a microsimulation of route selection.;A framework for handling dynamic activity planning, and the development of activity attribute planning models implemented within that framework, forms the core focus of the work. The main contribution of this model is in explicitly representing the dynamics of the activity planning and scheduling process, rather than modeling the observed activity-travel patterns as is usually done. The time-dependence in activity planning is represented through a set of planning constrained choice models, which limit decisions based on previous plans. Additionally, a new GPS-based travel survey was developed and implemented in order to capture data to allow the modeling of activity planning dynamics. This survey was a long-term prompted recall study which observed the planning behavior of 112 individuals over an average period of 14 days. The survey represents a significant new source of data on the planning dynamics of individuals and was used extensively in the ADAPTS model development.;The model was implemented for a baseline case for the Chicago region using recent survey data. The estimated individual activity pattern and region-level network characteristics were found to conform well to observations and compared favorably to the existing regional model. It is felt that the use of this type of model allows for a much wider range of advanced transportation demand management and environmental mitigation policies to be evaluated, especially those policies expected to appreciably change user behavior, such as flexible work scheduling, ridesharing, road pricing and many others.