SECTION 1.2 OPTMOV AND OPTOUT--ACCESS TO FINAL MATRICES

On completion of the optimization, two final options are available. By calling the subroutine OPTMOV (OPTOUT), any (or all) of the arrays used by the selected method which are of interest are available for use (output) by the main program. This option is utilized by placing one of the following statements after the call to subroutine OPT:

CALL OPTMOV(N,A,NSIZE) CALL OPTOUT(N)

where

N = 0 (OPTOUT only) for output of all available arrays

N = 1 for the H matrix (approx. inverse of the 2nd derivative).
in DFP or inverse 2nd derivative matrix from GRADX, NSIZE=NP*NP

N = 2 for the 2nd derivative at the optimum in GRADX. NSIZE=NP*NP

N = 3 for the gradient at the optimum in DFP or GRADX. NSIZE=NP

N = 4 for the simplex of points used by NMSIMP or the basis
vectors of POWELL. NSIZE=NP*NP

N = 5 for the final search direction. NSIZE=NP

N = 6 for the approx. diagonal of the second derivative in DFP.
NSIZE=NP

N = 7 for the covariance matrix (see Section
1.4)

NSIZE=NP*NP

A = For OPTMOV only, array or matrix into which the information
is to be put. A must be dimensioned with NP rows if it is to be
used as a two dimensional array.

NSIZE = for OPTMOV only, size of array to be transferred.

Return to

|Sect. 1.4|Sect. 10.1|Sect. 13.3|Beginning|