Hierarchical IPF: Generating a synthetic population for Switzerland

Abstract

Agent-based microsimulation models for land use or transportation simulate the behavior of agents over time, although at different time scales and with different goals.For both kinds of models, the initial step is the definition of agents and their relationships.Synthesizing the population of agents often is the only solution, due to privacy and cost constraints.In this paper, we assume that the model simulates persons grouped into households, and a person/household population needs to be synthesized.However, the methodology presented here can be applied to other kinds of agent relationships as well, e.g., persons and jobs/workplaces or persons and activity chains.Generating a synthetic population requires (a) reweighting of an initial population, taken from census or other survey data, with respect to current constraints, and (b) choosing the households that belong to the generated population.We propose an algorithm that estimates household-level weights that fit the control totals at both person and household levels.This eliminates the need to account for person-level control during the generation of synthetic households.The algorithm essentially performs a proportional fitting in the domains of both households and persons, and introduces an entropy-optimizing fitting step to switch between these two domains.We evaluate the performance of our algorithm by generating a synthetic population for Switzerland and checking it against the complete Swiss census.The validation is performed using informationbased and distance-based statistics, and proves that the new algorithm is highly competitive to the approaches presented by Ye et al. (2009) and Bar-Gera et al. (2009).

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 54 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.