A Framework for Validation of Computer Models

TLDR

Bayesian methods are well suited to addressing key validation challenges, such as quantifying multiple error sources, integrating diverse information, and updating assessments as new data become available, and they enable inference about predictive error in untested scenarios. The study presents a framework for evaluating computer models to determine whether they adequately represent reality. The framework consists of a six-step Bayesian and likelihood-based procedure, implemented in a resistance spot welding test bed to illustrate each step. The test bed application demonstrates the framework’s ability to contextualize each validation step. Keywords: Bayesian analysis, identifiability, model discrepancy, prediction.

Abstract

AbstractWe present a framework that enables computer model evaluation oriented toward answering the question: Does the computer model adequately represent reality? The proposed validation framework is a six-step procedure based on Bayesian and likelihood methodology. The Bayesian methodology is particularly well suited to treating the major issues associated with the validation process: quantifying multiple sources of error and uncertainty in computer models, combining multiple sources of information, and updating validation assessments as new information is acquired. Moreover, it allows inferential statements to be made about predictive error associated with model predictions in untested situations. The framework is implemented in a test bed example of resistance spot welding, to provide context for each of the six steps in the proposed validation process.KEY WORDS : Bayesian analysisIdentifiabilityModel discrepancyPrediction