A Theoretical Framework For Simulation Validation: Popper’S Falsifications

Abstract

This paper provides a philosophical yet scientific assessment of simulation validation. It suggests that the activity of simulation validation has firm theoretical roots in falsificationism, Karl Popper’s method of scientific discovery. Popper’s falsificationist view of science holds that a theory is scientific only if it is falsifiable; that is, if it makes specific predictions whose nonoccurrence will discredit it. According to Popper, while scientific theories can be falsified, they can never be proven true or verified. If a theory agrees with real world observations, it is still only, regarded as a tentative or provisional representation of reality.Simulationists are practitioners of falsificationism when they compare the output generated by a simulation model with the output data expected from or generated by a real world system. If the simulation model output data conflict with the real world output data, the simulation model is falsified and is therefore rejected. If the simulation model output data and the real world output data are consistent with each other, the simulation model is not rejected, but neither is it accepted or “proved true”. It is provisionally accepted as “valid” because it has not been falsified.The paper compares Popper’s approach with the traditional verificationist view of the scientific method and discusses the problems of induction and demarcation. Simulation is shown to be an empirical science that attempts to establish theoretical, predictive relationships which describe a real world system.