A Framework for Modeling Occupancy Schedules and Local Trips Based on Human Activity Patterns in the City

This paper presents a methodology for the generation of building occupancy schedules and trips in urban areas based on regional activity-based travel surveys. Using eigendecomposition-based Principal Component Analysis (PCA) and the k-means clustering algorithm, the method identifies activities as well as city resident types. The outputs are typical occupancy schedules for building performance simulation, and activity profiles for mobility modeling in urban-scale simulation. As a proof of concept the framework is applied to the 2010/2011 Massachusetts Department of Transportation (MassDOT) travel survey. Four clusters of occupants were identified, and an example building occupancy schedule and complementary activity distributions were determined. A discussion about how such activity profiles can be used to generate a set of user behavior profiles for neighborhood and city models is presented.