MathGroup Archive 2006

[Date Index] [Thread Index] [Author Index]

Search the Archive

Re: Trigonometric simplification

In article <edjgsd$lt0$1 at>,
 "Carl K. Woll" <carlw at> wrote:

> Yes, using ChebyshevT was what I had recommended before, and is a bit 
> simpler for this particular problem. See:
> The reason that I now suggest using TrigExpand and ArcCos is that this 
> method seems to be a bit more general. For example, how would you convert
> 2 + 3 Sin[a] + Sin[3a]
> into a sum of powers of Sin[a]? 

For odd multiples,

  Sin[(k_ /; OddQ[k]) a_] :> (-1)^((k - 1)/2) ChebyshevT[k, Sin[a]]

Similarly, for even multiples,

 Sin[(k_ /; EvenQ[k]) a_]:> (-1)^(k/2 + 1) Cos[a] ChebyshevU[k-1, Sin[a]]

> Or, how would you convert the following mixture of sines and cosines 
> into a trigonometric polynomial in Cos[a]:
> 2 + 3 Cos[a] Sin[a] + Cos[3a] Sin[3a]

It is _not_ a trigonometric polynomial in Cos[a]! Clearly it involves 
Sin[a]. Moreover, using

 TrigExpand[2 + 3 Cos[a] Sin[a] + Cos[3a] Sin[3a] /. a -> ArcCos[x]] /.
   x -> Cos[a] // Simplify

leads to

 2+4 Cos[a] Sqrt[Sin[a]^2]+Cos[3a] Sqrt[Sin[a]^2]+Cos[5a] Sqrt[Sin[a]^2]

with unnecessary Sqrts. The "standard" decomposition is

 2 + 3 Cos[a] Sin[a] + Cos[3a] Sin[3a] /.
  {Cos[n_Integer a_] :> ChebyshevT[n, Cos[a]],
   Sin[n_Integer a_] :> Sin[a] ChebyshevU[n-1, Cos[a]]}

> I think both techniques are good to have in one's toolbox.

Agreed. However, for trig polynomials, one only needs a small set of 
transformations involving Chebyshev polynomials, and these always lead 
to a valid representation.


Paul Abbott                                      Phone:  61 8 6488 2734
School of Physics, M013                            Fax: +61 8 6488 1014
The University of Western Australia         (CRICOS Provider No 00126G)    

  • Prev by Date: Re: Why doesn't Mathematica solve this simple differential equation?
  • Next by Date: Re: repost -Difficulties with xml
  • Previous by thread: Re: Re: Trigonometric simplification
  • Next by thread: difficulties with xml