integration of rational function of sine and cosine
The integration task
∫R(sinx,cosx)𝑑x, | (1) |
where the integrand is a rational function of sinx and cosx, changes via the Weierstrass substitution
tanx2=t | (2) |
to a form having an integrand that is a rational function of t. Namely, since x=2arctant, we have
dx= 2⋅11+t2dt, | (3) |
and we can substitute
sinx=2t1+t2,cosx=1-t21+t2, | (4) |
getting
∫R(sinx,cosx)𝑑x= 2∫R(2t1+t2,1-t21+t2)dt1+t2. |
Proof of the formulae (4): Using the double angle formulas of sine and cosine and then dividing the numerators and the denominators by cos2x2 we obtain
sinx=2sinx2cosx2sin2x2+cos2x2=2tanx21+tan2x2=2t1+t2, |
cosx=cos2x2-sin2x2sin2x2+cos2x2=1-tan2x21+tan2x2=1-t21+t2. |
Example. The above formulae give from ∫dxsinx the result
∫dxsinx=∫1+t22t⋅2⋅11+t2𝑑t=∫dtt=ln|t|+C=ln|tanx2|+C |
(which can also be expressed in the form -ln|cscx+cotx|+C; see the goniometric formulas).
Note 1. The substitution (2) is sometimes called the ‘‘universal trigonometric substitution’’ (http://planetmath.org/UniversalTrigonometricSubstitution). In practice, it often gives rational functions that are too complicated. In many cases, it is more profitable to use other substitutions:
-
•
In the case ∫R(sinx)cosxdx the substitution sinx=t is simpler.
-
•
Similarly, in the case ∫R(cosx)sinxdx the substitution cosx=t is simpler.
-
•
If the integrand depends only on tanx, the substitution tanx=t is simpler.
-
•
If the integrand is of the form R(sin2x,cos2x), one can use the substitution tanx=t; then
cos2x=11+tan2x=11+t2, sin2x=1-cos2x=t21+t2, dx=dt1+t2.
Example. The integration of ∫dxcos4x𝑑x is of the last case:
∫dxcos4x𝑑x=∫1(cos2x)2𝑑x=∫(1+t2)2⋅dt1+t2=∫(1+t2)𝑑t=t33+t+C=13tan3x+tanx+C. |
Example. The integral I=∫dxcos3x𝑑x=∫sec3xdx
is a peculiar case in which one does not have to use the substitutions mentioned above, as integration by parts is a simpler method for evaluating this integral. Thus,
u=secx⇒du=secxtanxdx;dv=sec2xdx⇒v=tanx. |
Therefore,
I=∫sec3xdx=secxtanx-∫secxtan2xdx=secxtanx-∫secx(sec2x-1)𝑑x=secxtanx-I+∫secxdx,
and consequently
∫dxcos3x𝑑x=12(secxtanx+ln|secx+tanx|)+C. |
Note 2. There is also the ‘‘universal hyperbolic substitution’’ for integrating rational functions of hyperbolic sine and cosine:
tanhx2=t,dx=2dt1-t2,sinhx=2t1-t2,coshx=1+t21-t2 |
References
- 1 Л. Д. Кдрячев: Математичецкии анализ. Издательство ‘‘ВүсшаяШкола’’. Москва (1970).
Title | integration of rational function of sine and cosine |
Canonical name | IntegrationOfRationalFunctionOfSineAndCosine |
Date of creation | 2013-03-22 17:05:15 |
Last modified on | 2013-03-22 17:05:15 |
Owner | pahio (2872) |
Last modified by | pahio (2872) |
Numerical id | 29 |
Author | pahio (2872) |
Entry type | Topic |
Classification | msc 26A36 |
Synonym | universal trigonometric substitution |
Related topic | GoniometricFormulae |
Related topic | SubstitutionForIntegration |
Related topic | WeierstrassSubstitutionFormulas |
Related topic | EulersSubstitutionsForIntegration |
Related topic | ErrorsCanCancelEachOtherOut |
Defines | universal hyperboloc substitution |