1 | ! MATH_LIB: Mathematics procedures for F90 |
---|
2 | ! Compiled/Modified: |
---|
3 | ! 07/01/06 John Haynes (haynes@atmos.colostate.edu) |
---|
4 | ! |
---|
5 | ! gamma (function) |
---|
6 | ! path_integral (function) |
---|
7 | ! avint (subroutine) |
---|
8 | |
---|
9 | module math_lib |
---|
10 | implicit none |
---|
11 | |
---|
12 | contains |
---|
13 | |
---|
14 | ! ---------------------------------------------------------------------------- |
---|
15 | ! function GAMMA |
---|
16 | ! ---------------------------------------------------------------------------- |
---|
17 | function gamma(x) |
---|
18 | implicit none |
---|
19 | ! |
---|
20 | ! Purpose: |
---|
21 | ! Returns the gamma function |
---|
22 | ! |
---|
23 | ! Input: |
---|
24 | ! [x] value to compute gamma function of |
---|
25 | ! |
---|
26 | ! Returns: |
---|
27 | ! gamma(x) |
---|
28 | ! |
---|
29 | ! Coded: |
---|
30 | ! 02/02/06 John Haynes (haynes@atmos.colostate.edu) |
---|
31 | ! (original code of unknown origin) |
---|
32 | |
---|
33 | ! ----- INPUTS ----- |
---|
34 | real*8, intent(in) :: x |
---|
35 | |
---|
36 | ! ----- OUTPUTS ----- |
---|
37 | real*8 :: gamma |
---|
38 | |
---|
39 | ! ----- INTERNAL ----- |
---|
40 | real*8 :: pi,ga,z,r,gr |
---|
41 | real*8 :: g(26) |
---|
42 | integer :: k,m1,m |
---|
43 | |
---|
44 | pi = acos(-1.) |
---|
45 | if (x ==int(x)) then |
---|
46 | if (x > 0.0) then |
---|
47 | ga=1.0 |
---|
48 | m1=x-1 |
---|
49 | do k=2,m1 |
---|
50 | ga=ga*k |
---|
51 | enddo |
---|
52 | else |
---|
53 | ga=1.0+300 |
---|
54 | endif |
---|
55 | else |
---|
56 | if (abs(x) > 1.0) then |
---|
57 | z=abs(x) |
---|
58 | m=int(z) |
---|
59 | r=1.0 |
---|
60 | do k=1,m |
---|
61 | r=r*(z-k) |
---|
62 | enddo |
---|
63 | z=z-m |
---|
64 | else |
---|
65 | z=x |
---|
66 | endif |
---|
67 | data g/1.0,0.5772156649015329, & |
---|
68 | -0.6558780715202538, -0.420026350340952d-1, & |
---|
69 | 0.1665386113822915,-.421977345555443d-1, & |
---|
70 | -.96219715278770d-2, .72189432466630d-2, & |
---|
71 | -.11651675918591d-2, -.2152416741149d-3, & |
---|
72 | .1280502823882d-3, -.201348547807d-4, & |
---|
73 | -.12504934821d-5, .11330272320d-5, & |
---|
74 | -.2056338417d-6, .61160950d-8, & |
---|
75 | .50020075d-8, -.11812746d-8, & |
---|
76 | .1043427d-9, .77823d-11, & |
---|
77 | -.36968d-11, .51d-12, & |
---|
78 | -.206d-13, -.54d-14, .14d-14, .1d-15/ |
---|
79 | gr=g(26) |
---|
80 | do k=25,1,-1 |
---|
81 | gr=gr*z+g(k) |
---|
82 | enddo |
---|
83 | ga=1.0/(gr*z) |
---|
84 | if (abs(x) > 1.0) then |
---|
85 | ga=ga*r |
---|
86 | if (x < 0.0) ga=-pi/(x*ga*sin(pi*x)) |
---|
87 | endif |
---|
88 | endif |
---|
89 | gamma = ga |
---|
90 | return |
---|
91 | end function gamma |
---|
92 | |
---|
93 | ! ---------------------------------------------------------------------------- |
---|
94 | ! function PATH_INTEGRAL |
---|
95 | ! ---------------------------------------------------------------------------- |
---|
96 | function path_integral(f,s,i1,i2) |
---|
97 | use m_mrgrnk |
---|
98 | use array_lib |
---|
99 | implicit none |
---|
100 | ! |
---|
101 | ! Purpose: |
---|
102 | ! evalues the integral (f ds) between f(index=i1) and f(index=i2) |
---|
103 | ! using the AVINT procedure |
---|
104 | ! |
---|
105 | ! Inputs: |
---|
106 | ! [f] functional values |
---|
107 | ! [s] abscissa values |
---|
108 | ! [i1] index of lower limit |
---|
109 | ! [i2] index of upper limit |
---|
110 | ! |
---|
111 | ! Returns: |
---|
112 | ! result of path integral |
---|
113 | ! |
---|
114 | ! Notes: |
---|
115 | ! [s] may be in forward or reverse numerical order |
---|
116 | ! |
---|
117 | ! Requires: |
---|
118 | ! mrgrnk package |
---|
119 | ! |
---|
120 | ! Created: |
---|
121 | ! 02/02/06 John Haynes (haynes@atmos.colostate.edu) |
---|
122 | |
---|
123 | ! ----- INPUTS ----- |
---|
124 | real*8, intent(in), dimension(:) :: f,s |
---|
125 | integer, intent(in) :: i1, i2 |
---|
126 | |
---|
127 | ! ---- OUTPUTS ----- |
---|
128 | real*8 :: path_integral |
---|
129 | |
---|
130 | ! ----- INTERNAL ----- |
---|
131 | real*8 :: sumo, deltah, val |
---|
132 | integer*4 :: nelm, j |
---|
133 | integer*4, dimension(i2-i1+1) :: idx |
---|
134 | real*8, dimension(i2-i1+1) :: f_rev, s_rev |
---|
135 | |
---|
136 | nelm = i2-i1+1 |
---|
137 | if (nelm > 3) then |
---|
138 | call mrgrnk(s(i1:i2),idx) |
---|
139 | s_rev = s(idx) |
---|
140 | f_rev = f(idx) |
---|
141 | call avint(f_rev(i1:i2),s_rev(i1:i2),(i2-i1+1), & |
---|
142 | s_rev(i1),s_rev(i2), val) |
---|
143 | path_integral = val |
---|
144 | else |
---|
145 | sumo = 0. |
---|
146 | do j=i1,i2 |
---|
147 | deltah = abs(s(i1+1)-s(i1)) |
---|
148 | sumo = sumo + f(j)*deltah |
---|
149 | enddo |
---|
150 | path_integral = sumo |
---|
151 | endif |
---|
152 | ! print *, sumo |
---|
153 | |
---|
154 | return |
---|
155 | end function path_integral |
---|
156 | |
---|
157 | ! ---------------------------------------------------------------------------- |
---|
158 | ! subroutine AVINT |
---|
159 | ! ---------------------------------------------------------------------------- |
---|
160 | subroutine avint ( ftab, xtab, ntab, a_in, b_in, result ) |
---|
161 | implicit none |
---|
162 | ! |
---|
163 | ! Purpose: |
---|
164 | ! estimate the integral of unevenly spaced data |
---|
165 | ! |
---|
166 | ! Inputs: |
---|
167 | ! [ftab] functional values |
---|
168 | ! [xtab] abscissa values |
---|
169 | ! [ntab] number of elements of [ftab] and [xtab] |
---|
170 | ! [a] lower limit of integration |
---|
171 | ! [b] upper limit of integration |
---|
172 | ! |
---|
173 | ! Outputs: |
---|
174 | ! [result] approximate value of integral |
---|
175 | ! |
---|
176 | ! Reference: |
---|
177 | ! From SLATEC libraries, in public domain |
---|
178 | ! |
---|
179 | !*********************************************************************** |
---|
180 | ! |
---|
181 | ! AVINT estimates the integral of unevenly spaced data. |
---|
182 | ! |
---|
183 | ! Discussion: |
---|
184 | ! |
---|
185 | ! The method uses overlapping parabolas and smoothing. |
---|
186 | ! |
---|
187 | ! Modified: |
---|
188 | ! |
---|
189 | ! 30 October 2000 |
---|
190 | ! 4 January 2008, A. Bodas-Salcedo. Error control for XTAB taken out of |
---|
191 | ! loop to allow vectorization. |
---|
192 | ! |
---|
193 | ! Reference: |
---|
194 | ! |
---|
195 | ! Philip Davis and Philip Rabinowitz, |
---|
196 | ! Methods of Numerical Integration, |
---|
197 | ! Blaisdell Publishing, 1967. |
---|
198 | ! |
---|
199 | ! P E Hennion, |
---|
200 | ! Algorithm 77, |
---|
201 | ! Interpolation, Differentiation and Integration, |
---|
202 | ! Communications of the Association for Computing Machinery, |
---|
203 | ! Volume 5, page 96, 1962. |
---|
204 | ! |
---|
205 | ! Parameters: |
---|
206 | ! |
---|
207 | ! Input, real ( kind = 8 ) FTAB(NTAB), the function values, |
---|
208 | ! FTAB(I) = F(XTAB(I)). |
---|
209 | ! |
---|
210 | ! Input, real ( kind = 8 ) XTAB(NTAB), the abscissas at which the |
---|
211 | ! function values are given. The XTAB's must be distinct |
---|
212 | ! and in ascending order. |
---|
213 | ! |
---|
214 | ! Input, integer NTAB, the number of entries in FTAB and |
---|
215 | ! XTAB. NTAB must be at least 3. |
---|
216 | ! |
---|
217 | ! Input, real ( kind = 8 ) A, the lower limit of integration. A should |
---|
218 | ! be, but need not be, near one endpoint of the interval |
---|
219 | ! (X(1), X(NTAB)). |
---|
220 | ! |
---|
221 | ! Input, real ( kind = 8 ) B, the upper limit of integration. B should |
---|
222 | ! be, but need not be, near one endpoint of the interval |
---|
223 | ! (X(1), X(NTAB)). |
---|
224 | ! |
---|
225 | ! Output, real ( kind = 8 ) RESULT, the approximate value of the integral. |
---|
226 | |
---|
227 | integer, intent(in) :: ntab |
---|
228 | |
---|
229 | integer,parameter :: KR8 = selected_real_kind(15,300) |
---|
230 | real ( kind = KR8 ), intent(in) :: a_in |
---|
231 | real ( kind = KR8 ) a |
---|
232 | real ( kind = KR8 ) atemp |
---|
233 | real ( kind = KR8 ), intent(in) :: b_in |
---|
234 | real ( kind = KR8 ) b |
---|
235 | real ( kind = KR8 ) btemp |
---|
236 | real ( kind = KR8 ) ca |
---|
237 | real ( kind = KR8 ) cb |
---|
238 | real ( kind = KR8 ) cc |
---|
239 | real ( kind = KR8 ) ctemp |
---|
240 | real ( kind = KR8 ), intent(in) :: ftab(ntab) |
---|
241 | integer i |
---|
242 | integer ihi |
---|
243 | integer ilo |
---|
244 | integer ind |
---|
245 | real ( kind = KR8 ), intent(out) :: result |
---|
246 | real ( kind = KR8 ) sum1 |
---|
247 | real ( kind = KR8 ) syl |
---|
248 | real ( kind = KR8 ) term1 |
---|
249 | real ( kind = KR8 ) term2 |
---|
250 | real ( kind = KR8 ) term3 |
---|
251 | real ( kind = KR8 ) x1 |
---|
252 | real ( kind = KR8 ) x2 |
---|
253 | real ( kind = KR8 ) x3 |
---|
254 | real ( kind = KR8 ), intent(in) :: xtab(ntab) |
---|
255 | logical lerror |
---|
256 | |
---|
257 | lerror = .false. |
---|
258 | a = a_in |
---|
259 | b = b_in |
---|
260 | |
---|
261 | if ( ntab < 3 ) then |
---|
262 | write ( *, '(a)' ) ' ' |
---|
263 | write ( *, '(a)' ) 'AVINT - Fatal error!' |
---|
264 | write ( *, '(a,i6)' ) ' NTAB is less than 3. NTAB = ', ntab |
---|
265 | stop |
---|
266 | end if |
---|
267 | |
---|
268 | do i = 2, ntab |
---|
269 | if ( xtab(i) <= xtab(i-1) ) then |
---|
270 | lerror = .true. |
---|
271 | exit |
---|
272 | end if |
---|
273 | end do |
---|
274 | |
---|
275 | if (lerror) then |
---|
276 | write ( *, '(a)' ) ' ' |
---|
277 | write ( *, '(a)' ) 'AVINT - Fatal error!' |
---|
278 | write ( *, '(a)' ) ' XTAB(I) is not greater than XTAB(I-1).' |
---|
279 | write ( *, '(a,i6)' ) ' Here, I = ', i |
---|
280 | write ( *, '(a,g14.6)' ) ' XTAB(I-1) = ', xtab(i-1) |
---|
281 | write ( *, '(a,g14.6)' ) ' XTAB(I) = ', xtab(i) |
---|
282 | stop |
---|
283 | end if |
---|
284 | |
---|
285 | result = 0.0D+00 |
---|
286 | |
---|
287 | if ( a == b ) then |
---|
288 | write ( *, '(a)' ) ' ' |
---|
289 | write ( *, '(a)' ) 'AVINT - Warning!' |
---|
290 | write ( *, '(a)' ) ' A = B, integral=0.' |
---|
291 | return |
---|
292 | end if |
---|
293 | ! |
---|
294 | ! If B < A, temporarily switch A and B, and store sign. |
---|
295 | ! |
---|
296 | if ( b < a ) then |
---|
297 | syl = b |
---|
298 | b = a |
---|
299 | a = syl |
---|
300 | ind = -1 |
---|
301 | else |
---|
302 | syl = a |
---|
303 | ind = 1 |
---|
304 | end if |
---|
305 | ! |
---|
306 | ! Bracket A and B between XTAB(ILO) and XTAB(IHI). |
---|
307 | ! |
---|
308 | ilo = 1 |
---|
309 | ihi = ntab |
---|
310 | |
---|
311 | do i = 1, ntab |
---|
312 | if ( a <= xtab(i) ) then |
---|
313 | exit |
---|
314 | end if |
---|
315 | ilo = ilo + 1 |
---|
316 | end do |
---|
317 | |
---|
318 | ilo = max ( 2, ilo ) |
---|
319 | ilo = min ( ilo, ntab - 1 ) |
---|
320 | |
---|
321 | do i = 1, ntab |
---|
322 | if ( xtab(i) <= b ) then |
---|
323 | exit |
---|
324 | end if |
---|
325 | ihi = ihi - 1 |
---|
326 | end do |
---|
327 | |
---|
328 | ihi = min ( ihi, ntab - 1 ) |
---|
329 | ihi = max ( ilo, ihi - 1 ) |
---|
330 | ! |
---|
331 | ! Carry out approximate integration from XTAB(ILO) to XTAB(IHI). |
---|
332 | ! |
---|
333 | sum1 = 0.0D+00 |
---|
334 | |
---|
335 | do i = ilo, ihi |
---|
336 | |
---|
337 | x1 = xtab(i-1) |
---|
338 | x2 = xtab(i) |
---|
339 | x3 = xtab(i+1) |
---|
340 | |
---|
341 | term1 = ftab(i-1) / ( ( x1 - x2 ) * ( x1 - x3 ) ) |
---|
342 | term2 = ftab(i) / ( ( x2 - x1 ) * ( x2 - x3 ) ) |
---|
343 | term3 = ftab(i+1) / ( ( x3 - x1 ) * ( x3 - x2 ) ) |
---|
344 | |
---|
345 | atemp = term1 + term2 + term3 |
---|
346 | |
---|
347 | btemp = - ( x2 + x3 ) * term1 & |
---|
348 | - ( x1 + x3 ) * term2 & |
---|
349 | - ( x1 + x2 ) * term3 |
---|
350 | |
---|
351 | ctemp = x2 * x3 * term1 + x1 * x3 * term2 + x1 * x2 * term3 |
---|
352 | |
---|
353 | if ( i <= ilo ) then |
---|
354 | ca = atemp |
---|
355 | cb = btemp |
---|
356 | cc = ctemp |
---|
357 | else |
---|
358 | ca = 0.5D+00 * ( atemp + ca ) |
---|
359 | cb = 0.5D+00 * ( btemp + cb ) |
---|
360 | cc = 0.5D+00 * ( ctemp + cc ) |
---|
361 | end if |
---|
362 | |
---|
363 | sum1 = sum1 & |
---|
364 | + ca * ( x2**3 - syl**3 ) / 3.0D+00 & |
---|
365 | + cb * 0.5D+00 * ( x2**2 - syl**2 ) & |
---|
366 | + cc * ( x2 - syl ) |
---|
367 | |
---|
368 | ca = atemp |
---|
369 | cb = btemp |
---|
370 | cc = ctemp |
---|
371 | |
---|
372 | syl = x2 |
---|
373 | |
---|
374 | end do |
---|
375 | |
---|
376 | result = sum1 & |
---|
377 | + ca * ( b**3 - syl**3 ) / 3.0D+00 & |
---|
378 | + cb * 0.5D+00 * ( b**2 - syl**2 ) & |
---|
379 | + cc * ( b - syl ) |
---|
380 | ! |
---|
381 | ! Restore original values of A and B, reverse sign of integral |
---|
382 | ! because of earlier switch. |
---|
383 | ! |
---|
384 | if ( ind /= 1 ) then |
---|
385 | ind = 1 |
---|
386 | syl = b |
---|
387 | b = a |
---|
388 | a = syl |
---|
389 | result = -result |
---|
390 | end if |
---|
391 | |
---|
392 | return |
---|
393 | end subroutine avint |
---|
394 | |
---|
395 | end module math_lib |
---|