Fidelity evaluation framework

While the modeling and simulation community commonly uses the word fidelity, there exists no clearly accepted definition or method of measuring fidelity. We make the following contributions. We present a new approach for measuring fidelity: the fidelity evaluation framework (FEF), that uses a referent, or a formal representation of reality that is intermediate between reality and the simulation. This foundation is advantageous because isolates subjectivity from the fidelity evaluation to well-defined framework components: development of the referent and assignment of weights to different referent components. We then provide the first example of the composition of a detailed referent and two models based on a real-world system with the FEF. We propose and illustrate three new methods of computing fidelity objectively within the FEF: category-based, model-based, and weight-based. This experiment proved that the FEF can provide meaningful and useful measurements of fidelity