derivnum

Function: derivnum
Section: sums
C-Name: derivnum0
Prototype: V=GEDGp
Help: derivnum(X=a,expr,{ind=1}): numerical derivation of expr with respect to
 X at X = a. The order of derivation is given by parameter 'ind', which can
 be a vector.
Wrapper: (,Gp)
Description:
  (gen,gen):gen:prec derivnum(${2 cookie}, ${2 wrapper}, $1, $prec)
  (gen,gen,gen):gen:prec derivfunk(${2 cookie}, ${2 wrapper}, $1, $3, $prec)
Doc: numerical derivation of \var{expr} with respect to $X$ at $X=a$. The
 order of derivation is 1 by default.

 \bprog
 ? derivnum(x=0, sin(exp(x))) - cos(1)
 %1 = 0.E-38
 @eprog
 A clumsier approach, which would not work in library mode, is
 \bprog
 ? f(x) = sin(exp(x))
 ? f'(0) - cos(1)
 %2 = 0.E-38
 @eprog

 \item When $a$ is a numerical type (integer, rational number, real number or
 \typ{COMPLEX} of such), performs numerical derivation.

 \item When $a$ is a (polynomial, rational function or) power series, compute
 \kbd{derivnum(t=a,f)} as $f'(a) = (f(a))'/a'$:
 \bprog
 ? derivnum(x = 1 + t, sqrt(x))
 %1 = 1/2 - 1/4*t + 3/16*t^2 - 5/32*t^3 + ... + O(t^16)
 ? derivnum(x = 1/(1 + t), sqrt(x))
 %2 = 1/2 + 1/4*t - 1/16*t^2 + 1/32*t^3 + ... + O(t^16)
 ? derivnum(x = 1 + t + O(t^17), sqrt(x))
 %3 = 1/2 - 1/4*t + 3/16*t^2 - 5/32*t^3 + ... + O(t^16)
 @eprog

 If the parameter \var{ind} is present, it can be

 \item a nonnegative integer $m$, in which case we return $f^{(m)}(x)$;

 \item or a vector of orders, in which case we return the vector of
 derivatives.

 \bprog
 ? derivnum(x = 0, exp(sin(x)), 16) \\ 16-th derivative
 %1 = -52635599.000000000000000000000000000000

 ? round( derivnum(x = 0, exp(sin(x)), [0..13]) )  \\ 0-13-th derivatives
 %2 = [1, 1, 1, 0, -3, -8, -3, 56, 217, 64, -2951, -12672, 5973, 309376]
 @eprog

 \synt{derivfunk}{void *E, GEN (*eval)(void*,GEN), GEN a, GEN ind, long prec}.
 Also available is
 \fun{GEN}{derivfun}{void *E, GEN (*eval)(void *, GEN), GEN a, long prec}.
 If $a$ is a numerical type (\typ{INT}, \typ{FRAC}, \typ{REAL} or
 \typ{COMPLEX} of such, we have
 \fun{GEN}{derivnumk}{void *E, GEN (*eval)(void *, GEN, long), GEN a, GEN ind, long prec}
 and
 \fun{GEN}{derivnum}{void *E, GEN (*eval)(void *, GEN, long prec), GEN a, long prec}

Function: _derivfun
Section: programming/internals
C-Name: derivfun0
Prototype: GGGD1,L,p
Help: _derivfun(args,def,closure,k) numerical kth-derivation of closure with respect to
 the first variable at args