Uncertainty Characterization for Social Simulation

Session chairs:

Vivek Srikrishnan, Cornell Atkinson Center for Sustainability, Cornell University

Tatiana Filatova, Delft University of Technology


Computational simulations of socioeconomic systems play an important role in understanding their dynamics and making projections of their future states. Simulation models are subject to many uncertainties, including choices about model structure, parameter values, and data. How these uncertainties are treated can have profound implications for inferences and projections. Uncertainty and sensitivity analysis methods have advanced significantly over the past years. Yet, not all of these techniques have been widely adopted by the ABM community. This uptake has been hindered by the multidimensionality of ABM output data and the complex adaptive nature of socioeconomic systems. This special track aims to renew a dialog in the ABM community on the best practices of uncertainty analysis. Specifically, we seek papers that explicitly address the role of uncertainty in social simulation, ranging from simulation experiments which characterize structural or parametric uncertainties to methodological advances enabling efficient global sensitivity analysis for ABMs.