Structural control analysis of simulation models

Abstract Modelers of socio-economic dynamics may fail to determine if a simulation is demonstrating effective system control; in particular, whether, and to what extent, each available policy handle (control input) can influence which system stocks. There lacks theoretical grounds for systematically investigating the control properties of simulation models in social sciences. To fill this gap, we adapt the structural controllability theory for artificial dynamic systems to simulation models while amplifying the classic theory with application contents based on graph theory analysis. Our efforts help (1) identify model variables that are control inputs; (2) on the model graph, examine the control capability of each identified control input: its ability to influence which system stocks; and (3) determine the controllability of each system stock: which control inputs influence it. This post-modeling workflow is summarized as the structural control analysis (SCA) for simulation models. Benefits of SCA are discussed through a series of examples.