Google
 

Trailing-Edge - PDP-10 Archives - decuslib10-02 - 43,50145/dse35.doc
There are 2 other files named dse35.doc in the archive. Click here to see a list.
SUBROUTINE DSE35

PURPOSE
   TO COMPUTE A VECTOR OF SMOOTHED FUNCTION VALUES GIVEN A
   VECTOR OF FUNCTION VALUES WHOSE ENTRIES CORRESPOND TO
   EQUIDISTANTLY SPACED ARGUMENT VALUES.

USAGE
   CALL DSE35(Y,Z,NDIM,IER)

DESCRIPTION OF PARAMETERS
   Y	 -  GIVEN VECTOR OF DOUBLE PRECISION FUNCTION VALUES
	    (DIMENSION NDIM)
   Z	 -  RESULTING VECTOR OF DOUBLE PRECISION SMOOTHED
	    FUNCTION VALUES (DIMENSION NDIM)
   NDIM  -  DIMENSION OF VECTORS Y AND Z
   IER	 -  RESULTING ERROR PARAMETER
	    IER = -1  - NDIM IS LESS THAN 5
	    IER =  0  - NO ERROR

REMARKS
   (1)	IF IER=-1 THERE HAS BEEN NO COMPUTATION.
   (2)	 Z CAN HAVE THE SAME STORAGE ALLOCATION AS Y.  IF Y IS
	 DISTINCT FROM Z, THEN IT IS NOT DESTROYED.

SUBROUTINE AND FUNCTION SUBPROGRAMS REQUIRED
   NONE

METHOD
   IF X IS THE (SUPPRESSED) VECTOR OF ARGUMENT VALUES, THEN
   EXCEPT AT THE POINTS X(1),X(2),X(NDIM-1) AND X(NDIM), EACH
   SMOOTHED VALUE Z(I) IS OBTAINED BY EVALUATING AT X(I) THE
   LEAST-SQUARES POLYNOMIAL OF DEGREE 3 RELEVANT TO THE 5
   SUCCESSIVE POINTS (X(I+K),Y(I+K)) K = -2,-1,...,2.  (SEE
   HILDEBRAND, F.B., INTRODUCTION TO NUMERICAL ANALYSIS,
   MC GRAW-HILL, NEW YORK/TORONTO/LONDON, 1956, PP. 295-302.)