• In this example it seems reasonable to run the model for the scenarios at the 4 extremes (corners) and the centre point for confirmation. • The results from these runs, which should be reported as confidence intervals, are signified by X and • By simple linear interpolation, the model user could then predict the results for the other four scenarios, denoted by E(X ) for estimated value. • This does of course assume that the solution space is linear and smooth over the region being investigated.