Google
 

Trailing-Edge - PDP-10 Archives - decuslib20-02 - decus/20-0026/dqatr.doc
There are 2 other files named dqatr.doc in the archive. Click here to see a list.
SUBROUTINE DQATR


PURPOSE
   TO COMPUTE AN APPROXIMATION FOR INTEGRAL(FCT(X), SUMMED
   OVER X FROM XL TO XU).

USAGE
   CALL DQATR (XL,XU,EPS,NDIM,FCT,Y,IER,AUX)
   PARAMETER FCT REQUIRES AN EXTERNAL STATEMENT.

DESCRIPTION OF PARAMETERS
   XL	  - DOUBLE PRECISION LOWER BOUND OF THE INTERVAL.
   XU	  - DOUBLE PRECISION UPPER BOUND OF THE INTERVAL.
   EPS	  - SINGLE PRECISION UPPER BOUND OF THE ABSOLUTE ERROR.
   NDIM   - THE DIMENSION OF THE AUXILIARY STORAGE ARRAY AUX.
	    NDIM-1 IS THE MAXIMAL NUMBER OF BISECTIONS OF
	    THE INTERVAL (XL,XU).
   FCT	  - THE NAME OF THE EXTERNAL DOUBLE PRECISION FUNCTION
	    SUBPROGRAM USED.
   Y	  - RESULTING DOUBLE PRECISION APPROXIMATION FOR THE
	    INTEGRAL VALUE.
   IER	  - A RESULTING ERROR PARAMETER.
   AUX	  - AUXILIARY DOUBLE PRECISION STORAGE ARRAY WITH
	    DIMENSION NDIM.

REMARKS
   ERROR PARAMETER IER IS CODED IN THE FOLLOWING FORM
   IER=0  - IT WAS POSSIBLE TO REACH THE REQUIRED ACCURACY.
	    NO ERROR.
   IER=1  - IT IS IMPOSSIBLE TO REACH THE REQUIRED ACCURACY
	    BECAUSE OF ROUNDING ERRORS.
   IER=2  - IT WAS IMPOSSIBLE TO CHECK ACCURACY BECAUSE NDIM
	    IS LESS THAN 5, OR THE REQUIRED ACCURACY COULD NOT
	    BE REACHED WITHIN NDIM-1 STEPS. NDIM SHOULD BE
	    INCREASED.

SUBROUTINES AND FUNCTION SUBPROGRAMS REQUIRED
   THE EXTERNAL DOUBLE PRECISION FUNCTION SUBPROGRAM FCT(X)
   MUST BE CODED BY THE USER. ITS DOUBLE PRECISION ARGUMENT X
   SHOULD NOT BE DESTROYED.

METHOD
   EVALUATION OF Y IS DONE BY MEANS OF TRAPEZOIDAL RULE IN
   CONNECTION WITH ROMBERGS PRINCIPLE. ON RETURN Y CONTAINS
   THE BEST POSSIBLE APPROXIMATION OF THE INTEGRAL VALUE AND
   VECTOR AUX THE UPWARD DIAGONAL OF ROMBERG SCHEME.
   COMPONENTS AUX(I) (I=1,2,...,IEND, WITH IEND LESS THAN OR
   EQUAL TO NDIM) BECOME APPROXIMATIONS TO INTEGRAL VALUE WITH
   DECREASING ACCURACY BY MULTIPLICATION WITH (XU-XL).
   FOR REFERENCE, SEE
   (1) FILIPPI, DAS VERFAHREN VON ROMBERG-STIEFEL-BAUER ALS
       SPEZIALFALL DES ALLGEMEINEN PRINZIPS VON RICHARDSON,
       MATHEMATIK-TECHNIK-WIRTSCHAFT, VOL.11, ISS.2 (1964),
       PP.49-54.
   (2) BAUER, ALGORITHM 60, CACM, VOL.4, ISS.6 (1961), PP.255.