[Population Modeling] Introducing PopGen

[Ram Pendyala]

Greetings:  I am writing this message to introduce the synthetic population generator called PopGen, and our research group is excited to be part of this list.

PopGen is a synthetic population generator that was developed in 2008 to support the development and implementation of activity-based microsimulation models for travel demand forecasting. The advent of activity-based microsimulation models of travel demand has provided the ability to simulate activity-travel patterns of individual travelers in time and space. The application of activity-based travel demand model systems requires the generation of synthetic populations whose attributes match those of the general population within small geographies (say, a traffic analysis zone, census tract, block group, or block).  In Ye et al (2009), we developed and introduced a heuristic iterative procedure known as the IPU (Iterative Proportional Updating) algorithm to facilitate the generation of a synthetic population that matches census data with respect to both household and person attributes. This algorithm has been implemented in PopGen, which is an open source synthetic population generator.  In PopGen, the iterative proportional fitting (IPF) procedure is applied first to both household- and person-level control variables of interest to obtain the number of households and persons in each cell of the respective joint distributions.  Appropriate rounding procedures are applied to obtain cell "constraints" that must be matched through the population synthesis process. The IPU algorithm computes weights for sample households such that household-level as well as person-level marginal distributions are matched as closely as possible.
In addition to developing the IPU algorithm, we also developed and implemented an entropy-maximization based approach to deriving weights for sample records such that both household and person attributes of interest are matched very closely.  A relaxed formulation of this algorithm accommodates for data inconsistencies, which is often encountered when dealing with input data used for transportation demand forecasting. The PopGen software package incorporates both the IPU and entropy-maximization methodologies and the user can choose the method to be applied in any specific PopGen run.  We have since enhanced the PopGen package to accommodate control variables at multiple geographic resolutions through an enhancement of the IPU algorithm.  Our current work focuses on the development of a full-fledged demographic and socio-economic evolution model system so that a base (current) year synthetic population can be aged through various lifecycle processes over time, thus enabling the generation of a future year synthetic population.  In addition, we are implementing a cloud-version of PopGen to enable ease of use.

The link for various PopGen related resources (including software packages) is: http://www.mobilityanalytics.org/popgen.html

The following publications offer details about the methods embedded in the PopGen software (and are available at the web link noted above).


1.       Konduri, K.C., D. You, V.M. Garikapati, and R.M. Pendyala (2016) Application of an Enhanced Population Synthesis Model that Accommodates Controls at Multiple Geographic Resolutions. Transportation Research Record, Journal of the Transportation Research Board (forthcoming).

2.       Ye, X., K. Konduri, R.M. Pendyala, B. Sana, and P. Waddell (2009) A Methodology to Match Distributions of Both Household and Person Attributes in the Generation of Synthetic Populations. Proceedings of 88th Annual Meeting of the Transportation Research Board,  National Research Council, Washington, D.C.

3.       Bar-Gera, H., K. Konduri, B. Sana, X. Ye, and R.M. Pendyala (2009) Estimating Survey Weights with Multiple Constraints Using Entropy Optimization Methods. Proceedings of 88th Annual Meeting of the Transportation Research Board,  National Research Council, Washington, D.C.

A  very preliminary initial attempt at the development of our demographic/population evolution modeling system is described in the following PhD dissertation. https://repository.asu.edu/attachments/135142/content/Paul_asu_0010E_14038.pdf

Thank you very much for bringing this list to our attention and we look forward to collaborations that will further advance the science of population modeling.

Best regards,
Ram


Ram M. Pendyala, PhD
Professor, Transportation Systems
School of Sustainable Engineering and the Built Environment
Arizona State University
Tempe, AZ 85281
Personal Website<http://rampendyala.weebly.com/>
Research Website<http://www.mobilityanalytics.org/>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://simtk.org/pipermail/popmodwkgrpimag-news/attachments/20170409/b8f2d0ff/attachment.html>