chfdv.f 5.3 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165
  1. *DECK CHFDV
  2. SUBROUTINE CHFDV (X1, X2, F1, F2, D1, D2, NE, XE, FE, DE, NEXT,
  3. + IERR)
  4. C***BEGIN PROLOGUE CHFDV
  5. C***PURPOSE Evaluate a cubic polynomial given in Hermite form and its
  6. C first derivative at an array of points. While designed for
  7. C use by PCHFD, it may be useful directly as an evaluator
  8. C for a piecewise cubic Hermite function in applications,
  9. C such as graphing, where the interval is known in advance.
  10. C If only function values are required, use CHFEV instead.
  11. C***LIBRARY SLATEC (PCHIP)
  12. C***CATEGORY E3, H1
  13. C***TYPE SINGLE PRECISION (CHFDV-S, DCHFDV-D)
  14. C***KEYWORDS CUBIC HERMITE DIFFERENTIATION, CUBIC HERMITE EVALUATION,
  15. C CUBIC POLYNOMIAL EVALUATION, PCHIP
  16. C***AUTHOR Fritsch, F. N., (LLNL)
  17. C Lawrence Livermore National Laboratory
  18. C P.O. Box 808 (L-316)
  19. C Livermore, CA 94550
  20. C FTS 532-4275, (510) 422-4275
  21. C***DESCRIPTION
  22. C
  23. C CHFDV: Cubic Hermite Function and Derivative Evaluator
  24. C
  25. C Evaluates the cubic polynomial determined by function values
  26. C F1,F2 and derivatives D1,D2 on interval (X1,X2), together with
  27. C its first derivative, at the points XE(J), J=1(1)NE.
  28. C
  29. C If only function values are required, use CHFEV, instead.
  30. C
  31. C ----------------------------------------------------------------------
  32. C
  33. C Calling sequence:
  34. C
  35. C INTEGER NE, NEXT(2), IERR
  36. C REAL X1, X2, F1, F2, D1, D2, XE(NE), FE(NE), DE(NE)
  37. C
  38. C CALL CHFDV (X1,X2, F1,F2, D1,D2, NE, XE, FE, DE, NEXT, IERR)
  39. C
  40. C Parameters:
  41. C
  42. C X1,X2 -- (input) endpoints of interval of definition of cubic.
  43. C (Error return if X1.EQ.X2 .)
  44. C
  45. C F1,F2 -- (input) values of function at X1 and X2, respectively.
  46. C
  47. C D1,D2 -- (input) values of derivative at X1 and X2, respectively.
  48. C
  49. C NE -- (input) number of evaluation points. (Error return if
  50. C NE.LT.1 .)
  51. C
  52. C XE -- (input) real array of points at which the functions are to
  53. C be evaluated. If any of the XE are outside the interval
  54. C [X1,X2], a warning error is returned in NEXT.
  55. C
  56. C FE -- (output) real array of values of the cubic function defined
  57. C by X1,X2, F1,F2, D1,D2 at the points XE.
  58. C
  59. C DE -- (output) real array of values of the first derivative of
  60. C the same function at the points XE.
  61. C
  62. C NEXT -- (output) integer array indicating number of extrapolation
  63. C points:
  64. C NEXT(1) = number of evaluation points to left of interval.
  65. C NEXT(2) = number of evaluation points to right of interval.
  66. C
  67. C IERR -- (output) error flag.
  68. C Normal return:
  69. C IERR = 0 (no errors).
  70. C "Recoverable" errors:
  71. C IERR = -1 if NE.LT.1 .
  72. C IERR = -2 if X1.EQ.X2 .
  73. C (Output arrays have not been changed in either case.)
  74. C
  75. C***REFERENCES (NONE)
  76. C***ROUTINES CALLED XERMSG
  77. C***REVISION HISTORY (YYMMDD)
  78. C 811019 DATE WRITTEN
  79. C 820803 Minor cosmetic changes for release 1.
  80. C 890411 Added SAVE statements (Vers. 3.2).
  81. C 890531 Changed all specific intrinsics to generic. (WRB)
  82. C 890831 Modified array declarations. (WRB)
  83. C 890831 REVISION DATE from Version 3.2
  84. C 891214 Prologue converted to Version 4.0 format. (BAB)
  85. C 900315 CALLs to XERROR changed to CALLs to XERMSG. (THJ)
  86. C***END PROLOGUE CHFDV
  87. C Programming notes:
  88. C
  89. C To produce a double precision version, simply:
  90. C a. Change CHFDV to DCHFDV wherever it occurs,
  91. C b. Change the real declaration to double precision, and
  92. C c. Change the constant ZERO to double precision.
  93. C
  94. C DECLARE ARGUMENTS.
  95. C
  96. INTEGER NE, NEXT(2), IERR
  97. REAL X1, X2, F1, F2, D1, D2, XE(*), FE(*), DE(*)
  98. C
  99. C DECLARE LOCAL VARIABLES.
  100. C
  101. INTEGER I
  102. REAL C2, C2T2, C3, C3T3, DEL1, DEL2, DELTA, H, X, XMI, XMA, ZERO
  103. SAVE ZERO
  104. DATA ZERO /0./
  105. C
  106. C VALIDITY-CHECK ARGUMENTS.
  107. C
  108. C***FIRST EXECUTABLE STATEMENT CHFDV
  109. IF (NE .LT. 1) GO TO 5001
  110. H = X2 - X1
  111. IF (H .EQ. ZERO) GO TO 5002
  112. C
  113. C INITIALIZE.
  114. C
  115. IERR = 0
  116. NEXT(1) = 0
  117. NEXT(2) = 0
  118. XMI = MIN(ZERO, H)
  119. XMA = MAX(ZERO, H)
  120. C
  121. C COMPUTE CUBIC COEFFICIENTS (EXPANDED ABOUT X1).
  122. C
  123. DELTA = (F2 - F1)/H
  124. DEL1 = (D1 - DELTA)/H
  125. DEL2 = (D2 - DELTA)/H
  126. C (DELTA IS NO LONGER NEEDED.)
  127. C2 = -(DEL1+DEL1 + DEL2)
  128. C2T2 = C2 + C2
  129. C3 = (DEL1 + DEL2)/H
  130. C (H, DEL1 AND DEL2 ARE NO LONGER NEEDED.)
  131. C3T3 = C3+C3+C3
  132. C
  133. C EVALUATION LOOP.
  134. C
  135. DO 500 I = 1, NE
  136. X = XE(I) - X1
  137. FE(I) = F1 + X*(D1 + X*(C2 + X*C3))
  138. DE(I) = D1 + X*(C2T2 + X*C3T3)
  139. C COUNT EXTRAPOLATION POINTS.
  140. IF ( X.LT.XMI ) NEXT(1) = NEXT(1) + 1
  141. IF ( X.GT.XMA ) NEXT(2) = NEXT(2) + 1
  142. C (NOTE REDUNDANCY--IF EITHER CONDITION IS TRUE, OTHER IS FALSE.)
  143. 500 CONTINUE
  144. C
  145. C NORMAL RETURN.
  146. C
  147. RETURN
  148. C
  149. C ERROR RETURNS.
  150. C
  151. 5001 CONTINUE
  152. C NE.LT.1 RETURN.
  153. IERR = -1
  154. CALL XERMSG ('SLATEC', 'CHFDV',
  155. + 'NUMBER OF EVALUATION POINTS LESS THAN ONE', IERR, 1)
  156. RETURN
  157. C
  158. 5002 CONTINUE
  159. C X1.EQ.X2 RETURN.
  160. IERR = -2
  161. CALL XERMSG ('SLATEC', 'CHFDV', 'INTERVAL ENDPOINTS EQUAL', IERR,
  162. + 1)
  163. RETURN
  164. C------------- LAST LINE OF CHFDV FOLLOWS ------------------------------
  165. END