The activity-based model (ABM) is an alternative approach for demand modeling that represents travel demand in a region as a result of many individual travel decisions.
In classical macroscopic demand models, the population is aggregated into behaviorally homogeneous person groups. For each group, activity pairs or activity chains are modeled. The calculation of the demand is based on negative utilities of traveling. These models are also referred to as trip-based or tour-based demand models, because they focus on the pure number of trips or chains rather than on the formation of trips in the context of persons and their individual attributes. The results of a macroscopic demand model are several trip matrices differentiated by person group, trip purpose, and mode.
In contrast, in activity-based models (ABM) the focus lies on individual persons and their mobility. Activity-based models are microscopic demand models in which mobility decisions of all persons are individually simulated as successive discrete choices. Such a choice can be a long-term decision such as the choice of workplace or a short-term decision such as the number of tours and trips. Activity-based models are popular for some areas of demand modeling, as they model the behavior of individuals in a very intuitive way and have the potential to better account for some aspects of mobility (e.g. new modes, multi-modality, household interactions,…) than aggregated models.
The individuals are represented as a "synthetic population". The persons in a synthetic population are not exact reproductions of the true population, but are typically generated by Monte Carlo simulations of person attributes (gender, age, income, job, home location,…) based on statistical demographic data. The decisions of the persons depend on the characteristics associated with the generated persons. As a result of the model, schedules are generated in the form of sequences of activity executions. The schedules contain information about the activities, for example start times, duration and locations, as well as the trips including the choice of mode.
While the general concepts of activity-based models are similar, no common structure has emerged that is used as a generally accepted standard among transport planners. Therefore, activity-based models may differ in many respects. This concerns the definitions of the utility functions, the decisions considered and their order, the decision models themselves, as well as various model parameters. The approach of microscopic demand modeling in PTV Visum is therefore not to provide a complete activity-based model implementation, but to provide useful data structures, data management tools, COM APIs, file input/output, and visualizations for creating and applying custom activity-based models. The integration of custom activity-based models in PTV Visum allows you to benefit from all other features such as handling of large data sets, model sequence control, powerful assignments, and evaluation tools.
The following figure shows the basic features of the concept.
The synthetic population is calculated externally or can be generated in PTV Visum, for example on the basis of survey data. The data of the persons and households are stored in the new data structures persons and households(Managing persons and Managing households). Like other activity locations, households are also located by means of coordinates. All objects of the ABM data model can be extended by user-defined attributes.
The calculation of schedules and the resulting tours and trips (Managing tours and Managing trips) is based on scripts that must be created by the user as part of the modeling process. The results are saved as schedules, activity executions, tours, and trips. Lists are available for editing and evaluating these object types. They are synchronized with the network editor, which allows geographical examination of a single tour, as shown in the following figure.
|
Note: In the directory C:\Users\Public\Documents\PTV Vision\PTV Visum 2022/Examples, you can find an example of use on this topic. The Demand ABM example explains the structure of an activity-based model in PTV Visum. |
