Microsimulation system for predicting leisure activity-travel patterns

Development of a model of annual activity-travel patterns of leisure and vacation travel is reported. The simulation system, called Merlin, is a hybrid model system consisting of discrete choice models and rule-based models. It predicts the annual number of day trips and vacations, and the profile of such trips and vacations, including duration, travel party, timing, destination, type of accommodation, transport mode, day of departure, and expenditures. Separate models are derived for each of these facets. Microsimulation is used to predict aggregate patterns. The model's performance is satisfactory.