random-distributions/gamma.ss
;;; PLT Scheme Science Collection
;;; random-distributions/gamma.ss
;;; Copyright (c) 2004-2006 M. Douglas Williams
;;;
;;; This library is free software; you can redistribute it and/or
;;; modify it under the terms of the GNU Lesser General Public
;;; License as published by the Free Software Foundation; either
;;; version 2.1 of the License, or (at your option) any later version.
;;;
;;; This library is distributed in the hope that it will be useful,
;;; but WITHOUT ANY WARRANTY; without even the implied warranty of
;;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
;;; Lesser General Public License for more details.
;;;
;;; You should have received a copy of the GNU Lesser General Public
;;; License along with this library; if not, write to the Free
;;; Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
;;; 02111-1307 USA.
;;;
;;; -------------------------------------------------------------------
;;;
;;; This module implements the gamma distribution.  It is based on the
;;; Random Number Distributions in the GNU Scientific Library.
;;;
;;; Version  Date      Description
;;; 1.0.0    09/28/04  Marked as ready for Release 1.0.  Added
;;;                    contracts for functions.  (Doug Williams)
;;; 1.1.0   02/09/06   Added cdf.  (Doug Williams)

(module gamma mzscheme
  
  (require (lib "contract.ss"))
  
  (provide/contract
   (random-gamma
    (case-> (-> random-source? (>/c 0.0) real? (>=/c 0.0))
            (-> (>/c 0.0) real? (>=/c 0.0))))
   (random-gamma-int
    (-> random-source? natural-number/c (>=/c 0.0)))
   (gamma-pdf
    (-> real? (>/c 0.0) real? (>=/c 0.0)))
   (gamma-cdf
    (-> real? (>/c 0.0) real? (real-in 0.0 1.0))))
  
  (require "../machine.ss")
  (require "../math.ss")
  (require "../random-source.ss")
  (require "../special-functions/gamma.ss")
  
  ;; random-gamma: random-source x real x real -> real
  ;; random-gamma: real x real -> real
  ;;
  ;; This function returns a random variate from the gamma
  ;; distribution with parameters a and b.
  (define random-gamma
    (case-lambda
      ((r a b)
       (let ((na (inexact->exact (floor a))))
         (cond ((= a na)
                (* b (random-gamma-int r na)))
               ((= na 0.0)
                (* b (random-gamma-frac r a)))
               (else
                (* b (+ (random-gamma-int r na)
                        (random-gamma-frac r (- a na))))))))
      ((a b)
       (random-gamma (current-random-source) a b))))
  
  ;; random-gamma-int: random-source x integer -> real
  (define (random-gamma-int r na)
    (if (< na 12)
        ;; Note: for 12 iterations we are safe against underflow,
        ;; since the smallest positive random number is O(2^-32).
        ;; This means the smallest product is 2^(12*32) = 10^-116,
        ;; which is in the range of double precision.
        (let ((prod 1.0))
          (do ((i 0 (+ i 1)))
              ((= i na) (- (log prod)))
            (set! prod (* prod (random-uniform r)))))
        (random-gamma-large r na)))
  
  ;; random-gamma-large: random-source x real -> real
  (define (random-gamma-large r a)
    (let ((sqa (sqrt (- (* 2.0 a) 1.0)))
          (x 0.0)
          (y 0.0)
          (v 0.0))
      (let loop1 ()
        (let loop2 ()
          (set! y (tan (* pi (random-uniform r))))
          (set! x (+ (* sqa y) a -1))
          (if (<= x 0.0) (loop2)))
        (set! v (random-uniform r))
        (if (> v (- (* (+ 1.0 (* y y)) 
                       (exp (- (* (- a 1.0) (log (/ x (- a 1.0))))
                               (* sqa y))))))
            (loop1)))
      x))
 
  ;; random-gamma-frac: random-source x real -> real
  (define (random-gamma-frac r a)
    (let ((p (/ e (+ a e)))
          (q 0.0)
          (x 0.0)
          (u 0.0)
          (v 0.0))
      (let loop ()
        (set! u (random-uniform r))
        (set! v (random-uniform r))
        (if (< u p)
            (begin
              (set! x (exp (* (/ 1.0 a) (log v))))
              (set! q (exp (- x))))
            (begin
              (set! x (- 1.0 (log v)))
              (set! q (exp (* (- a 1.0) (log x))))))
        (if (>= (random-uniform r) q)
            (loop)))
      x))
  
  ;; gamma-pdf: real x real x real -> real
  ;;
  ;; This function computes the probability density p(x) for a gamma
  ;; function with parameters a and b.
  (define (gamma-pdf x a b)
    (cond ((< x 0)
           0.0)
          ((= x 0.0)
           (if (= a 1.0)
               (/ 1.0 b)
               0.0))
          ((= a 1.0)
           (exp (/ (/ (- x) b) b)))
          (else
           (/ (exp (+ (* (- a 1.0) (log (/ x b)))
                      (- (/ x b))
                      (- (lngamma a))))
              b))))
  
  ;; cdf
  
  (define LARGE-A 85)
  
  (define (norm-arg x a)
    (let ((t 0.0)
          (arg (+ (x (/ 1.0 3.0) (- a) (- (/ 0.02 a)))))
          (u (/ (- a 0.5) x)))
      (cond ((< (abs (- u 1.0)) double-epsilon)
             (set! t 0.0))
            ((< (abs u) double-epsilon)
             (set! t 1.0))
            ((> u 0.0)
             (let ((v (- 1.0 u)))
               (set! t (/ (+ 1.0 (- (* u u)) (* 2.0 u (log u))) (* v v)))))
            (else
             (set! t +nan.0)))
      (* arg (sqrt (/ (+ 1.0 t) x)))))
  
  ;; Wrapper for functions that do the work.
  (define (gamma-cdf x a b)
    (cond ((< x 0.0)
           0.0)
          (else
           (let ((y (/ x b)))
             (gamma-inc-P a y)))))
  
)