In Search of Sound 'in silico' Research - Validating a Complex System Simulation

Abstract

In logic, an argument is valid if true premises cannot be associated with false conclusions. However, this criterion is hardly applicable to real-world complex systems, hence it needs to be weakened: critical systems engineering, for example, requires acceptability rather than truth. Validity cannot be assumed, nor is it transitive; it has to be demonstrated within a context and related to a specific set of requirements. We present here work directed towards developing a structured argument of validity for a complex systems simulation - part of a plant ecology case-study. Firstly, we define the difficulties of claiming validity for complex systems simulations. We then proceed to analysing the spectrum of assumptions and abstractions that were involved in constructing the biological model and simulation, and how these influence the scope of our validity claim. Finally, we provide an initial structured argument for the simulation's validity, using critical systems techniques.