A simulation model without statistical analysis is useless (?!)

Simulation models of different types are commonly used in a wide variety of applications, such as crop yield prediction, forecasting climate change effects, quantifying effects of shocks on ecosystems or sociotechnical systems like water systems or electrical grids, or social simulation.

The utility of models for these applications depends on our ability to analyze their output, for instance, to get estimates of uncertainties in predictions, and to calibrate them to existing data. If we cannot analyze or calibrate models, we cannot validate or perhaps even understand their output, and then what is their use besides academic exercise?

For Agent Based Models and many other simulation model types there are no standardized analysis approaches. This hampers the utility of these models for engineering and policy applications. This project is aimed at the development and application of (preferably automated) statistical methodologies for model analysis and calibration, for example, the use of different types of sensitivity analysis or a sound design for numerical experiments.

Suggested starting literature:

Cariboni, J., et al. "The role of sensitivity analysis in ecological modelling." Ecological modelling 203.1-2 (2007): 167-182.

Lee, Ju Sung, et al. "The complexities of agent-based modeling output analysis." The journal of artificial societies and social simulation 18.4 (2015).

Ten Broeke, G., G. van Voorn, and A. Ligtenberg. "Which sensitivity analysis method should I use for my agent-based model?." Journal of Artificial Societies and Social Simulation 19.1 (2016).

Used skills

Statistics, software programming, some modelling when desired


A solid statistical background (with adequate grades for statistical courses) and knowledge of software programming (like R) is highly desirable.Affinity with simulation models or particular model applications will be helpful.