Mercurial > hg > octave-kai
changeset 7507:bc6573d2fa40
legendre.m: Added normalization and improved stability.
author | Ben Abbott <bpabbott@mac.com> |
---|---|
date | Wed, 20 Feb 2008 21:47:39 -0500 |
parents | 798b0a00e80c |
children | f501b22c0394 |
files | scripts/ChangeLog scripts/specfun/legendre.m |
diffstat | 2 files changed, 149 insertions(+), 66 deletions(-) [+] |
line wrap: on
line diff
--- a/scripts/ChangeLog +++ b/scripts/ChangeLog @@ -1,3 +1,8 @@ +2008-02-20 Marco Caliari <marco.caliari@univr.it> + + * specfun/legendre.m: Accept normalization options ("sch", "norm"), + and improve stability for higher orders. + 2008-02-20 John W. Eaton <jwe@octave.org> * strings/strcat.m: Detect cellstr args.
--- a/scripts/specfun/legendre.m +++ b/scripts/specfun/legendre.m @@ -1,4 +1,5 @@ ## Copyright (C) 2000, 2006, 2007 Kai Habel +## Copyright (C) 2008 Marco Caliari ## ## This file is part of Octave. ## @@ -17,33 +18,38 @@ ## <http://www.gnu.org/licenses/>. ## -*- texinfo -*- -## @deftypefn {Function File} {@var{L} =} legendre (@var{n}, @var{X}) +## @deftypefn {Function File} {@var{l} =} legendre (@var{n}, @var{x}) +## @deftypefnx {Function File} {@var{l} =} legendre (@var{n}, @var{x}, "unnorm") +## Compute the Legendre function of degree @var{n} and order @var{m} +## where all values for @var{m} = 0 @dots{} @var{n} are returned. +## @var{n} must be a non-negative scalar integer in the range greater +## than or equal to 0. The return value has one dimension more than @var{x}. ## -## Legendre Function of degree n and order m -## where all values for m = 0..@var{n} are returned. -## @var{n} must be a scalar in the range [0..255]. -## The return value has one dimension more than @var{x}. +## The Legendre Function of degree @var{n} and order @var{m}: ## ## @example -## The Legendre Function of degree n and order m -## ## @group ## m m 2 m/2 d^m ## P(x) = (-1) * (1-x ) * ---- P (x) ## n dx^m n ## @end group +## @end example ## -## with: -## Legendre polynomial of degree n +## @noindent +## with Legendre polynomial of degree @var{n}: ## +## @example ## @group ## 1 d^n 2 n -## P (x) = ------ [----(x - 1) ] +## P (x) = ------ [----(x - 1) ] ## n 2^n n! dx^n ## @end group +## @end example ## -## legendre(3,[-1.0 -0.9 -0.8]) returns the matrix +## @noindent +## @code{legendre (3, [-1.0, -0.9, -0.8])} returns the matrix: ## +## @example ## @group ## x | -1.0 | -0.9 | -0.8 ## ------------------------------------ @@ -53,80 +59,152 @@ ## m=3 | 0.00000 | -1.24229 | -3.24000 ## @end group ## @end example +## +## @deftypefnx {Function File} {@var{l} =} legendre (@var{n}, @var{x}, "sch") +## Compute the Schmidt semi-normalized associated Legendre function. +## The Schmidt semi-normalized associated Legendre function is related +## to the unnormalized Legendre functions by the following: +## +## For Legendre functions of degree n and order 0: +## +## @example +## @group +## 0 0 +## SP (x) = P (x) +## n n +## @end group +## @end example +## +## For Legendre functions of degree n and order m: +## +## @example +## @group +## m m m 2(n-m)! 0.5 +## SP (x) = P (x) * (-1) * [-------] +## n n (n+m)! +## @end group +## @end example +## +## @deftypefnx {Function File} {@var{l} =} legendre (@var{n}, @var{x}, "norm") +## Compute the fully normalized associated Legendre function. +## The fully normalized associated Legendre function is related +## to the unnormalized Legendre functions by the following: +## +## For Legendre functions of degree @var{n} and order @var{m} +## +## @example +## @group +## m m m (n+0.5)(n-m)! 0.5 +## NP (x) = P (x) * (-1) * [-------------] +## n n (n+m)! +## @end group +## @end example ## @end deftypefn -## FIXME Add Schmidt semi-normalized and fully normalized legendre functions - -## Author: Kai Habel <kai.habel@gmx.de> +## Author: Marco Caliari <marco.caliari@univr.it> -function L = legendre (n, x) +function retval = legendre (n, x, normalization) - warning ("legendre is unstable for higher orders"); - - if (nargin != 2) + if (nargin < 2 || nargin > 3) print_usage (); endif - if (! isscalar (n) || n < 0 || n > 255 || n != fix (n)) - error ("n must be a integer between 0 and 255]"); - endif + if (nargin == 3) + normalization = lower (normalization); + else + normalization = "unnorm"; + endif - if (! isvector (x) || any (x < -1 || x > 1)) - error ("x must be vector in range -1 <= x <= 1"); - endif + if (! isscalar (n) || n < 0 || n != fix (n)) + error ("legendre: n must be a non-negative scalar integer"); + endif - if (n == 0) - L = ones (size (x)); - elseif (n == 1) - L = [x; -sqrt(1 - x .^ 2)]; - else - i = 0:n; - a = (-1) .^ i .* bincoeff (n, i); - p = [a; zeros(size (a))]; - p = p(:); - p(length (p)) = []; - #p contains the polynom (x^2-1)^n + if (! isvector (x) || any (x < -1 || x > 1)) + error ("legendre: x must be vector in range -1 <= x <= 1"); + endif + + switch (normalization) + case "norm" + scale = sqrt (n+0.5); + case "sch" + scale = sqrt (2); + case "unnorm" + scale = 1; + otherwise + error ("legendre: expecting normalization option to be \"norm\", \"sch\", or \"unnorm\""); + endswitch - #now create a vector with 1/(2.^n*n!)*(d/dx).^n - d = [((n + rem(n, 2)):-1:(rem (n, 2) + 1)); 2 * ones(fix (n / 2), n)]; - d = cumsum (d); - d = fliplr (prod (d')); - d = [d; zeros(1, length (d))]; - d = d(1:n + 1) ./ (2 ^ n *prod (1:n)); - - Lp = d' .* p(1:length (d)); - #Lp contains the Legendre Polynom of degree n + ## Based on the recurrence relation below + ## m m m + ## (n-m+1) * P (x) = (2*n+1)*x*P (x) - (n+1)*P (x) + ## n+1 n n-1 + ## http://en.wikipedia.org/wiki/Associated_Legendre_function - # now create a polynom matrix with d/dx^m for m=0..n - d2 = flipud (triu (ones (n))); - d2 = cumsum (d2); - d2 = fliplr (cumprod (flipud (d2))); - d3 = fliplr (triu (ones (n + 1))); - d3(2:n + 1, 1:n) = d2; + for m = 1:n + lpm1 = scale; + lpm2 = (2*m-1) .* x .* scale; + lpm3 = lpm2; + for k = m+1:n + lpm3 = ((2*k-1) .* x .* lpm2 - (k+m-2) .* lpm1)/(k-m+1); + lpm1 = lpm2; + lpm2 = lpm3; + endfor + retval(m,:) = lpm3; + if (strcmp (normalization, "unnorm")) + scale = -scale * (2*m-1); + else + ## normalization == "sch" or normalization == "norm" + scale = scale / sqrt ((n-m+1)*(n+m))*(2*m-1); + endif + scale = scale .* sqrt(1-x.^2); + endfor - # multiply for each m (d/dx)^m with Lp(n,x) - # and evaluate at x - Y = zeros(n + 1, length (x)); - [dr, dc] = size (d3); - for m = 0:dr - 1 - Y(m + 1, :) = polyval (d3(m + 1, 1:(dc - m)) .* Lp(1:(dc - m))', x)(:)'; - endfor + retval(n+1,:) = scale; - # calculate (-1)^m*(1-x^2)^(m/2) for m=0..n at x - # and multiply with (d/dx)^m(Pnx) - l = length (x); - X = kron ((1 - x(:) .^ 2)', ones (n + 1, 1)); - M = kron ((0:n)', ones (1, l)); - L = X .^ (M / 2) .* (-1) .^ M .* Y; - endif + if (strcmp (normalization, "sch")) + retval(1,:) = retval(1,:) / sqrt (2); + endif + endfunction %!test -%! result=legendre(3,[-1.0 -0.9 -0.8]); +%! result = legendre (3, [-1.0 -0.9 -0.8]); %! expected = [ %! -1.00000 -0.47250 -0.08000 %! 0.00000 -1.99420 -1.98000 %! 0.00000 -2.56500 -4.32000 %! 0.00000 -1.24229 -3.24000 %! ]; -%! assert(result,expected,1e-5); +%! assert (result, expected, 1e-5); + +%!test +%! result = legendre (3, [-1.0 -0.9 -0.8], "sch"); +%! expected = [ +%! -1.00000 -0.47250 -0.08000 +%! 0.00000 0.81413 0.80833 +%! -0.00000 -0.33114 -0.55771 +%! 0.00000 0.06547 0.17076 +%! ]; +%! assert (result, expected, 1e-5); + +%!test +%! result = legendre (3, [-1.0 -0.9 -0.8], "norm"); +%! expected = [ +%! -1.87083 -0.88397 -0.14967 +%! 0.00000 1.07699 1.06932 +%! -0.00000 -0.43806 -0.73778 +%! 0.00000 0.08661 0.22590 +%! ]; +%! assert (result, expected, 1e-5); + +%!test +%! result = legendre (151, 0); +%! ## Don't compare to "-Inf" since it would fail on 64 bit systems. +%! assert (result(end) < -1.7976e308 && all (isfinite (result(1:end-1)))); + +%!test +%! result = legendre (150, 0); +%! ## This agrees with Matlab's result. +%! assert (result(end), 3.7532741115719e+306, 0.0000000000001e+306) + +