* ************************************************************************* subroutine rg( n, ind, d, h ) * ************************************************************************* * Purpose : * --------- * This routine sets the direction D to the steepest * descent direction, i.e. to minus the reduced gradient. * The components of these vectors that are considered * are the components of indices IND(i) for i=1 up to * i=n. * Parameters : * ------------ * n ( int ) * input : number of components of D to assign * to minus the reduced gradient value. * It is also the length of array IND. * output : unmodified. * ind ( int ) * input : array of length N, containing the indices * of the components of D to assign to minus * the reduced gradient value. * output : unmodified. * d ( dble ) * input : vector that is designed to contain the search * direction. * output : identical to the input value, except for * the components of indices IND(i) for i=1 * up to i=n, that are set to minus the * reduced gradient. * h ( dble ) * input : the reduced gradient vector. * output : unmodified. * Routines used : * --------------- * None. * Programming : * ------------- * D. Tuyttens * ======================================================================== * Routine parameters integer n, ind(*) double precision d(*), h(*) * Internal variables integer i, k do 10 i = 1 , n k = ind(i) d(k) = -h(k) 10 continue * return end