Preliminary Assessment of Activity Analysis and Modeling for Homeland Security Applications

Abstract

A review of 46 activity-based models is conducted and details features that are required for homeland security applications. Two examples of homeland security studies are included to highlight these needs. This review demonstrates that only a few models have the desired characteristics of fine spatial and temporal resolutions. Because the criticality of locations varies because of time or day of the week or both, depending on the presence of individuals and the fact that potential targets are specific locations within a city, data needs for modeling these types of scenarios exceed typical planning and forecasting modeling as well as research modeling. A discussion of a homeland security–related modeling application conducted by Los Alamos National Laboratory using TRANSIMS for the U.S. Department of Homeland Security is included to demonstrate the level of detail that is achieved by using nationally available data sources. In addition, the literature contains many data and information gaps, including virtually nonexistent models of security; identifying and inventorying residential locations; including more time constraints in models; modeling short-term changes in land use; carefully identifying and filling in missing subpopulations in surveys; accounting for fleets, services, and goods in models; better modeling of inter- and intrahousehold interactions, including weekly and seasonal variations in travel behavior; and developing a new paradigm for activity scheduling in panic and emergency situations.