Am Montag, den 08.06.2009, 22:42 +0200 schrieb Dieter Kaiser:
> Am Montag, den 08.06.2009, 13:17 -0400 schrieb Dan Gildea:
> > > From: Dieter Kaiser <drdieterkaiser at web.de>
> > > Subject: New definite integrals and new problems
> > >
> > > (%i58) integrate(exp(-x)*log(x),x,0,inf);
> > > (%o58) -'limit(-%e^-x*log(x)-gamma_incomplete(0,x),x,0,plus)
> > >
> > > The last step is to get the limit of the above expression which is known
> > > to be -%gamma.
> >
> > To get this limit, we probably need asymptotic expansions of the
> > special functions: see e.g. p 80 of Gruntz's thesis.
>
> I have got a paper from Richard Fateman - again many thanks - which
> gives three more general algorithm for integrals like the above. The
> algorithm are based on the differentiation of the integral
> representation of the Gamma, Beta and Incomplete Gamma functions.
>
> I think we have all functionality which is needed to implement these
> algorithm.
I have implemented the first algorithm with the help of the paper which
I have got from Richard Fateman to get the desired integral:
(%i62) ?defint\-log\-exp(exp(-t)*log(t),t);
(%o62) - %gamma
I have called the new Lisp routine directly to show the first result:
(defun m2-log-exp-1 (expr)
(when *debug-defint-log*
(format t "~&M2-LOG-EXP-1 with ~A~%" expr))
(m2 expr
'((mtimes)
(c freevar)
((mexpt) (z varp) (w freevar))
((mexpt) $%e ((coefft) (s freevar) (z varp)))
((mexpt) ((%log) (z varp)) (m freevar)))
nil))
(defun defint-log-exp (expr var)
(let ((x nil))
(cond
((setq x (m2-log-exp-1 expr))
(let ((c (cdras 'c x))
(w (cdras 'w x))
(m (cdras 'm x))
(s (cdras 's x)))
(setq s (mul -1 s))
(mul
(simplify (list '(%signum) s))
(power s (mul -1 (add m 1)))
($at ($diff (list '(%gamma) var) var m)
(list '(mequal)
var
(div (add w 1) s))))))
(t nil))))
This algorithm works for integrals of the type
integrate(t^w * log(t)^w * exp(-t^s), t, 0, inf)
This is the general solution:
(%i8) ?defint\-log\-exp(c*t^w*exp(-s*t)*log(t)^m,t);
(%o8) s^(-m-1)*signum(s)*(at('diff(gamma(t),t,m),t = (w+1)/s))
This is the first part of a more general algorithm. Furthermore, there
are two more algorithm for integrals with log and exp functions in the
paper.
I will have a look at defint to include this algorithm.
Dieter Kaiser