This project aims to provide algorithms for determining parameter estimates and associated uncertainties from aggregate data taking into account inconsistencies in the data.

The body of knowledge about complex systems involving a large number of parameters is comprised of measurement data and physical theory. Often a measurement result does not directly provide an estimate of a parameter value but instead that of a quantity, related through theory in a linear or nonlinear way, to a number of parameters. A parameter may figure in a number of experiments associated with different aspects of the physical theory. Each such experiment provides information about the parameter value, and an adjustment of the value of the parameter should take into account all relevant experimental evidence. This project will address the following questions:

• How do we assess if the data and physical theory are consistent with each other?
  
  
• If there is inconsistency, is there enough information (e.g. through multiple experiments) to indicate where it arises?
  
  
• How can the adjustment model be used to predict the likely impact of a proposed experiment or new item of information?

The methodologies developed will be tested on two metrology case studies: inter-laboratory comparisons and the adjustment of the fundamental constants.