SPG

Synthetic Population Generator

Location	File Name	Description
scenario_name\outputs\t*	householdsByHHCategory.csv	SPG1 modelwide summary of households by category
scenario_name\outputs\t*	spg2out_hh.csv	SPG2 Synthetic Population Household temporary (selected location fields for all HH Ids)
scenario_name\outputs\t*	SynPopH.csv	SPG2 Synthetic Population household attributes (current year study area resident households)
scenario_name\outputs\t*	SynPopP.csv	SPG2 Synthetic Population person attributes (current year study area resident persons)
scenario_name\outputs\t*	spg1.properties	Module-specific properties from current model run
scenario_name\outputs\t*	spg2.properties	Module-specific properties from current model run

SPG1

The SPG1 module is called via the com.pb.tlumip.spg.SPGnew class’s spg1 method. This method is actually called twice, once without the person-age constraint enabled to get an estimate for the total regional population (the module requires households be forecast, but not the actual population), and then a second time with the person-age constraint using the estimated population from the first run.

SPG2

The SPG2 module is called via the com.pb.tlumip.spg.SPGnew class’s spg2 method.

Method

SPG has 3 steps:

1. Balance - Multidimensional HH distribution in each TAZ usually using an IPF (Fratar-type) matrix balancing procedure
2. Discretize - List of HHs with controlled variables in each TAZ
3. Draw - Randomly join HHs from PUMS by controlled variables

Recent update for person age controls:

1. Convert person age controls to household controls:

     Hhs with 1 person age 34-45 for example
     Hhs with 2 persons age 34-45 for example
     …

1. Run SPG without person controls to get approximate persons
2. Allocate approximate persons to person controls using base year person age distributions
3. Re-run SPG with person controls and weight 1 person controls * 1, 2 person controls * 2, etc to account for person weighting Discretize and draws HHs to match results

Reference

Beckman, R.J., Baggerly, K.A., and McKay, M.D., 1996. Creating synthetic baseline populations. Transportation Research Part A, 30(6), 415-429.