Incorporating In-Home Activities into an Agent-Based Dynamic Activity Planning and Travel Simulation Model

This thesis addresses one of the important yet neglected areas in Activity-based Travel Demand models: In-home activities. In doing so, it attempts to extend the previously developed activity based framework called Agent-based Dynamic Activity Planning and Scheduling (ADAPTS) by integrating in-home activity models. The models are developed to capture the interdependencies between in-home and out-of-home activities while preserving the main dynamic planning structure of out-of home activities in ADAPTS. Additionally, the model components are designed so as to make non-transportation demand modeling applications feasible. This research focuses on generation of individuals' In-home and out-of-home activities as the simulation time runs. Simultaneously, individuals update their schedule based on activities they execute during the day and engage in trips accordingly. The link between in-home and out-of-home activities is implemented through a combination of rule-based and econometric models. Time of day sensitive activity type and duration models are proposed and implemented within the framework with the help of discrete choice, hazard-based and pairwise modeling concepts. This large-scale package could eventually be used for disaggregate demand forecasting purposes and targeted policies can be tested through relevant scenarios.