A framework to explore uncertainty in complex systems modeling

Keywords: uncertainty analysis, sensitivity analysis, complex systems modeling, modeling theory
Abstract Type: Virtual Paper Abstract

Abstract

Comprehensive representations of complex spatial systems usually require coupling models from various disciplines into a large consolidated model characterized by high computational cost, algorithmic complexity, and confounded uncertainty. The latter is usually perceived as an inconvenience that should be curtailed. It can be argued, however, that, if managed systematically, uncertainty can provide an opportunity to discover robust and sustainable futures. To this end, we propose three intertwined design principles that should guide the development of policy-relevant models: legitimacy, parsimony, and practicality. Model legitimacy encompasses the various sources of uncertainty – from expert opinion, through measurement error, to stakeholder preferences. Legitimacy aims at the faithful representation of the perspectives of all involved stakeholders (uncertainty differentiation). Model parsimony is necessary because legitimate models often result in overlapping system representations, which can be further simplified and grouped to minimize model complexity (uncertainty reduction). Finally, to satisfy practicality, we need to maintain a certain level of variability in models for experimentation that can augment consensus building (uncertainty exploration). These principles provide a foundation for a unique framework for studying complex systems that acknowledges inherent system uncertainty and aims at its exploitation.