SECTION 1.4 SUM-OF-SQUARES MINIMIZATION: ASYMPTOTIC COVARIANCE MATRICES
If the problem is a sum-of-squares minimization problem, the asymptotic covariance matrix corresponding to additive normal errors can be produced. This option is available only with GRADX and DFP. To use this option, the main program must include
COMMON/BNOBS/NOBS NOBS=i5
where i5 is the number of observations. The user may then call OPTOUT or OPTMOV, with N = 0 or 7. See Section 1.2. Setting NOBS to zero turns the option off.