Generation of Synthetic Populations in Social Simulations: A Review of Methods and Practices

With the aim of building realistic model of social systems, designers of agent-based models tend to incorporate more and more data and this data is of course having an impact on their outcomes. Among these data, those concerning the attributes of social agents, which together compose synthetic populations, are particularly important, but are usually difficult to collect and therefore to use in simulations. In this paper, we review the state of the art of methodologies and theories for building realistic synthetic populations for agent-based simulation models, as well as practices in the field of social simulation. We highlight, in particular, the discrepancies between theory and practice and outline the challenges in bridging this gap using a quantitative and narrative review of work published in JASSS between 2011 and 2021. Finally, we presents some ideas that can help modelers adopt better practices regarding synthetic population generation.