Robust population synthesis: a framework for reliability

Abstract

This paper proposes a novel framework for evaluating the reliability of synthetic population generation in agent-based transportation simulations (ABM-TOS). Our methodology categorizes key parameters into distinct categories, including sociodemographic attributes, household features, and travel patterns, to test them against the real data for a comprehensive validation of the generated population. The framework’s effectiveness is demonstrated by generating synthetic populations of varying sizes, representing 1% to 30% of Luxembourg’s 2021 population. The study showcases the importance of accurate population synthesis in ABM-TOS and highlights potential areas for improvement in future research.