
Does Boost.Math contain functions for Chebyshev polynomials? Preferably a recurrent chebyshev_next(). I can find functions for Legendre, Laguerre, and Hermite polynomials, but not Chebyshev.

On 13 May 2017, at 10:35, Bjorn Reese via Boost
wrote: Does Boost.Math contain functions for Chebyshev polynomials? Preferably a recurrent chebyshev_next().
I can find functions for Legendre, Laguerre, and Hermite polynomials, but not Chebyshev.
I second that request. Chebyshev polynomials are extremely useful in numerical applications: - evaluation does not loose precision even if the polynomial order is huge (order 1000 is no problem) - coefficients can be computed in O(N log(N)) with a FFT Best regards, Hans

On 13/05/2017 09:35, Bjorn Reese via Boost wrote:
Does Boost.Math contain functions for Chebyshev polynomials? Preferably a recurrent chebyshev_next().
No it doesn't as yet. However, the recurrence relation is trivial, likewise evaluation of an arbitrary Tn(x) in terms of trig functions? Even the roots have simple formulae. What did you want to do? John.
I can find functions for Legendre, Laguerre, and Hermite polynomials, but not Chebyshev.
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
--- This email has been checked for viruses by AVG. http://www.avg.com

On 05/15/2017 02:29 PM, John Maddock via Boost wrote:
However, the recurrence relation is trivial, likewise evaluation of an arbitrary Tn(x) in terms of trig functions? Even the roots have simple formulae.
They are indeed trivial, and I already have my own implementation. I just thought it might be nice to have in Boost.Math as we already have support for other polynomials.
What did you want to do?
I am using it for function approximation: https://en.wikipedia.org/wiki/Approximation_theory#Chebyshev_approximation
participants (3)
-
Bjorn Reese
-
Hans Dembinski
-
John Maddock