This procedure is used to create a synthetic population. Base households are duplicated at a suitable point in the model area in such a way that specified constraints are adhered to as far as possible across different geographical divisions (ABM generation of a synthetic population).
Prerequisites
The base households must exist in the model. Data on households is typically available as the result of a household survey as *.csv files, separated into households and persons. To import them into the model as demand data files *.dmd , their format only needs to be slightly adjusted (Importing predefined demand models *.dmd). You can then import the associated trips by importing synthetic population data (Import of synthetic population data).
|
Note: When importing, each household must be linked to a location. This location does not play a role in the further course, as the final location is the result of the generation of the synthetic population. It therefore makes sense to locate all base households at the same location. |
1. On the toolbar, click the 
Open 'Procedure sequence' window button.
The Procedure sequence window opens.
2. Add the ABM population synthesizer procedure of the Demand model category to the position of your choice in the procedure sequence (Setting up and starting the procedure sequence).
3. Make sure the added procedure is marked. Then, in the Operations section, click the Edit button.
4. Make the desired changes on the Basic settings page.
|
Element |
Description |
|
Filter for sample households |
Filter formula for the sample households. Determines which households from the survey sample are included. Only filtered households are given initial weights, scaled to target values (hard constraints), distributed across levels and multiplied. If no entry is made, the entire sample is used. |
|
Initial weights |
Initial weight per sample household. Basis for scaling to hard constraints and subsequent distribution across all levels. Use 1 if there is no pre-weighting of the sample. Note The information only controls the initial distribution. The subsequent balancing procedure adjusts the weights to the target values (hard constraints). |
5. Make the desired changes on the Geography levels page.
|
Element |
Description |
|
Network object type |
Specify, from higher-level to more detailed, on which levels constraints apply. Possible polygonal network object types:
Locations are always present as the most detailed level, cannot be edited, and must exist as the last row, as balancing takes place last at this level. If required, add rows using the Add button. Each level may only occur once. |
|
Network object filter |
Optional attribute formula to restrict the objects of the level |
6. Make the desired changes on the Constraints tab.
|
Element |
Description |
|
Network object type |
Selection of the level at which the constraint applies Notes When defining the constraints, the target value, an attribute at the objects of a geographic level, is compared with a household attribute or a household formula. Here you define the target values for each geography level. The target values of the constraints are compared with actual values as the sum of household attributes. For example, the target value for the population is compared with the actual value as the total number of people in the households. In addition to household attributes, personal attributes can also be used, such as the number of employees. The corresponding household attribute would then be a formula such as [No:Persons([Employed]=1)]. |
|
Target value attribute |
Object attribute of the selected level that provides the target value (e.g. population, number of cars) Note In addition to constraints, average values can also be defined as target values, such as the mean monthly household income per zone. The target value is then the sum of the incomes of all households in the zones. To determine this, the target value for the median income must be multiplied by the target value for the number of households. |
|
Formula for household value |
Specification of the household value, which is compared with the target value Note Please ensure that the values of the household attributes are not too high, i.e. do not exceed a value of around 20 or only very rarely. Otherwise, the calculation of the synthetic population can be time-consuming and numerically unstable. However, if the target values are significantly greater, the values must be scaled accordingly. The household income could then be given in whole thousands, for example, so an income of 3 would correspond to 3000 euros. |
7. Confirm with OK.
8. Execute the procedure.
The synthetic population is created.
