Population Synthesis Using Iterative Proportional Fitting (IPF): A Review and Future Research

Abstract Activity-Based travel demand modeling requires the detailed socioeconomic data of the study area population. Since the collection of such detailed data for the whole population is too expensive, if not infeasible, population synthesis has been proposed to predict the data and produce them synthetically based on a sample. This much cheaper alternative for forecasting population characteristics is based on iterative proportional fitting (IPF) and has been the focus of much research due to its many advantages. This paper seeks to critically review the state of the art of IPF, classify the peer-reviewed literature, investigate the major problems of IPF, and identify gaps for future research to enhance the usefulness of IPF for the synthesis. Our review shows that integer conversion and zero-cell are among the most important problems necessitating empirical investigation. Unbiased tabular (controlled) rounding methods should be developed to integerize the fractional numbers estimated by IPF for the frequency of household types. Zero-cell problem, although already dealt with, still lacks unbiased solutions. Simulation-based synthesis can help to avoid zero-cell problem while it has many other advantages.