Kirill Müller

Synthetic Population Generation by Combining a Hierarchical, Simulation-Based Approach with Reweighting by Generalized Raking

A recent approach for generating populations of synthetic individuals through simulation is extended to produce households of grouped individuals. The contingency tables of the generated populations match external controls on the individual and household levels while exhibiting far greater variety in composition than existing approaches can offer. The method involves a two-step approach. The first consists of a procedure based on Gibbs sampling, which has only recently been applied to population generation in transportation modeling and is generically called Markov chain Monte Carlo (MCMC). For this work, the model was generalized, and an extension was developed, hierarchical MCMC, which was able to generate a hierarchical structure. The second step, a postprocessing step, uses generalized raking (GR), which reweights the output from hierarchical MCMC to perfectly satisfy known marginal control totals on the individual and household levels. The application input data—a demographic sample and some known marginals from Singapore—added further complexities to the problem, which had not yet been explored in the current literature. Despite data challenges, consecutively applying the methods above produced realistic synthetic populations. Results confirm their goodness of fit and their generated hierarchical structures.

Using Survey Calibration and Statistical Matching to Reweight and Distribute Activity Schedules

The processes used to generate a 106-member agent population for a longrange study of 2030 travel in Switzerland are presented in this paper. This study was part of an effort to assess the effects of electric vehicles on the energy production and stability of the electric supply network. The process used well-established statistical methods—survey calibration and statistical matching. Both methods are described, and consistency with known approaches in transportation planning is shown. The paper introduces a new approach that allows exogenous specification of shares of activity types while maintaining the representativeness of the population: survey calibration is applied to satisfy these constraints, and statistical matching allows the joining of data sets with common variables. The discussion of the results for Switzerland focuses on the quality metrics available and highlights the links between the activity schedules and total shares of the activity types. Furthermore, the error introduced by the calibration and matching stages was analyzed and quantified, with special emphasis placed on uncontrolled sociodemographic and travel variables. The specified activity shares could be replicated almost perfectly; the resultant mean error in the uncontrolled variables was within the range of a few percentage points. Therefore, this approach is a viable alternative to a complicated estimation of an activity schedule model that would be necessary otherwise. Finally, two practical issues—data sets describing different populations and biased surveys—and their effect on the outcome were tested by suitably adjusting the input data.