Constraining Values in the Fit

It is possible to constrain values in the fit, for example backgrounds which have been measured independently. To do this the constraint flag (fifth column in the MXFF bank specifications section) must be set to 1 or 2. If the flag is set to 1 the two values required to constrain the distribution are the mean and standard deviation respectively and the constraint takes on a gaussian form. If the flag is set to 2, the two values are the minimum and maximum values for that distribution, though this section of the code is at present fundamentally flawed! The constraint on each distribution is calculated in mxf_constrain_parray.for for a given data event. This value is set to 1 if no constraint is declared. The product of these constraints is then incorporated in the likelihood function.