Measuring similarity for validation of computational electromagnetic modelling

Computational electromagnetics has been widely used since the ascendancy of computing. Validation of the resulting models has been a constant theme for developers, vendors and users, each wanting to be as confident as possible that the technique, code or model they are working with provides acceptable results in the situation in which it is being applied. However, no standardized procedure has been accepted to support this activity. A current IEEE activity aims to address this need, particularly with reference to applications in electromagnetic compatibility. One substantial question is how can confidence be quantified, allowing models to be compared with measurements to determine accuracy, or to compare models against other models to determine the level of difference. This paper addresses the issue of quantifying confidence setting this topic firmly in the progress of the standard activity to date.