Advances in population synthesis: fitting many attributes per agent and fitting to household and person margins simultaneously
Agent-based microsimulation models of transportation, land use or other socioeconomic processes require an initial synthetic population derived from census data, conventionally created using the iterative proportional fitting (IPF) procedure. This paper introduces a novel computational method that allows the synthesis of many more attributes and finer attribute categories than previous approaches, both of which are long-standing limitations discussed in the literature. Additionally, a new approach is used to fit household and person zonal attribute distributions simultaneously. This technique was first adopted to address limitations specific to Canadian census data, but could also be useful in U.S. and other applications. The results of each new method are evaluated empirically in terms of goodness-of-fit.
KeywordsIterative proportional fitting Population synthesis Microsimulation Agent-based Census microdata Transportation models Trip forecasting
This research was supported by funding from an Ontario Graduate Scholarship, the Transportation Association of Canada, and a Transport Canada Transportation Planning and Modal Integration grant. The authors would also like to thank Laine Ruus of the University of Toronto Data Library for her invaluable assistance.
- Arentze, T.A., Timmermans, H.J.: ALBATROSS Version 2: A Learning-Based Transportation Oriented Simulation System, Chapter 2. Eindhoven University of Technology, Eindhoven (2005)Google Scholar
- Barrett, C., et al: TRANSIMS 3.0 volume 3 (modules), Chapter 2 (population synthesizer). Unclassified Report LA-UR-00-1725. Los Alamos National Laboratories, Los Alamos (2003)Google Scholar
- Bowman, J.L.: A comparison of population synthesizers used in microsimulation models of activity and travel demand. Unpublished working paper. http://jbowman.net/papers/2004.Bowman.Comparison_of_PopSyns.pdf (2004)
- Guan, J.J.: Synthesizing family relationships between individuals for the ILUTE micro-simulation model. Bachelor’s Thesis, Department of Civil Engineering, University of Toronto (2002)Google Scholar
- Huang, Z., Williamson, P.: Comparison of synthetic reconstruction and combinatorial optimisation approaches to the creation of small-area microdata. Working Paper 2001/2. Department of Geography, University of Liverpool (2001)Google Scholar
- Pritchard, D.R.: Synthesizing agents and relationships for land use/transportation modelling. Master’s Thesis, Department of Civil Engineering, University of Toronto (2008)Google Scholar
- Wickens, T.D.: Multiway Contingency Tables Analysis for the Social Sciences. Lawrence Erlbaum Associates, Hillsdale (1989)Google Scholar
- Ye, X., Konduri, K., Pendyala, R.M., Sana, B., Waddell, P.: A methodology to match distributions of both household and person attributes in the generation of synthetic populations. Presented at the 88th Annual Meeting of the Transportation Research Board, Washington (2009)Google Scholar