The generation of a synthetic population is comparable to the weighting of a household survey: The basis is the results of a usual household survey with a path diary. Similar to aggregate demand models, the household survey does not have to be from the planning area but can be adjusted accordingly using socio-economic data from the planning area.
The synthetic population is formed from duplicates of the survey households such that the marginal socio-economic distributions of the generated population match those in the planning area. However, unlike conventional synthetic populations, here the schedule skeletons are duplicated in addition to the household and person attributes.
Sometimes household and person data should not be taken from a household survey because more accurate data is available from another source. In this case, suitable schedule skeletons from the household survey must be transferred to the synthetic persons. The data can be transferred to a scoring system, for example, in which certain household and person attributes (number of employees, number of children, employment, etc.) of the persons surveyed are compared with those of the synthetic population. A synthetic person then always receives the schedule skeleton of the person with the highest score.
A schedule skeleton can consist of several tours. The trip chains are known for each tour, furthermore the time durations of the activity executions and the start time of the main activity. The central activity is chosen as the main activity and the destination and mode choice for the entire tour is based on this.
The idea of a main activity is not a new construct: in tour-based demand models (e.g. the tour-based model), one main activity must be chosen for each trip chain, thus enabling "rubberbanding". The intermediate destinations are integrated into the outward and return journeys from work with as little detour as possible.
The spatial resolution at which the socio-economic data is used is not predefined. It depends on the data situation and may be available at different levels.
There are numerous software programs for generating a synthetic population. We recommend the use of PopulationSim, an open source software from a US consortium.
To install the software and create a synthetic population, please read the "Demand ActivitySim" example installed with the software in the folder C:\Users\Public\Documents\PTV Vision\PTV Visum 2025\Examples\Demand ActivitySim, which contains a detailed description.
|
Note: The PopulationSim software is developed in the U.S. as part of the ActivitySim project by a consortium of planning organizations and departments of transportation; see also https://activitysim.github.io/. |
