Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

While solving trigonometric problems I've noticed possible pattern that reminded me of Fibonacci numbers and Pascal triangle. So I tried to find next "element" of this pattern (8 degree exponent ) and failed to do so. So can someone please tell me is there a pattern or just "coincidence". Possible pattern :

$\sin^2x + \cos^2x = 1 $

$\sin^4x + \cos^4x = 1 - 2\sin^2x\cos^2x$

$\sin^6x + \cos^6x = 1 - 3\sin^2x\cos^2x$

share|improve this question
    
I can tell you that there is a pattern :) –  Ahmed Masud Dec 31 '11 at 14:23
add comment

1 Answer

up vote 8 down vote accepted

There is a pattern but I do not think it is as simple as what the first three terms might suggest.

Let me use the abbreviation $s := \sin x$ and $c := \cos x$. In this notation, we have $$ (s^6 + c^6) = (s^6 + c^6)(s^2 + c^2) = (s^8 + c^8) + s^2c^2(s^4+c^4). $$ Now plugging in the expressions you obtained for $s^6+c^6$ and $s^4+c^4$, we get $$ 1 - 3s^2c^2 = (s^8+c^8)+s^2c^2(1-2s^2c^2), $$ or: $$ s^8+ c^8 = 1 - 4s^2c^2 + 2s^4c^4. $$


Alternate derivation: Since $s^2 + c^2 = 1$, let $s^2 = \frac{1}{2} + u$ and $c^2 = \frac12 - u$ for some $u$. Then $$ s^2 c^2 = \frac14 - u^2. \tag{$\dagger$} $$ Now calculating $s^8+c^8$ is straightforward, if a bit involved: $$ \begin{align*} s^8 + c^8 &= \left(\frac12 + u\right)^4 + \left(\frac12 - u\right)^4 \\\\ &= 2 \left( \frac{1}{2^4} + u^4 + 6 \cdot \frac{1}{2^2} \cdot u^2 \right) \\\\ &= \frac{1}{8} +2 u^4 + 3 u^2 \\\\ &\stackrel{(\dagger)}{=} \frac{1}{8} +2 \left( \frac14 - s^2 c^2 \right)^2 + 3 \left( \frac14 - s^2 c^2 \right) \\\\ &= \frac18 + 2\cdot \frac{1}{4^2} + \frac34 +2 s^4 c^4 - 2 \cdot 2 \cdot \frac14 \cdot s^2 c^2 - 3 s^2 c^2 \\\\ &= 1 + 2 s^4 c^4 - 4s^2 c^2 . \end{align*} $$

Although this seems longer, the second method generalises easily for $s^{2n}+c^{2n}$ for all $n$.


What does $\sin^{2n} x + \cos^{2n} x$ look like?

It is a standard fact that any symmetric polynomial in $\lambda_1$ and $\lambda_2$ can be expressed in terms of the elementary symmetric polynomials, which, in this case, is just $\lambda_1 + \lambda_2$ and $\lambda_1 \lambda_2$. Now for $\lambda_1 = s^2 = \sin^2 x$ and $\lambda_2 = c^2 = \cos^2 x$, we have $\lambda_1 + \lambda_2 = 1$. That is, every symmetric polynomial in $s = \sin x$ and $c = \cos x$ with the restriction that every term has an even degree in both variables can be expressed as a polynomial in $\sin x\cos x$.

For the special case of $s^{2n} + c^{2n}$, I already explained how to get such a representation: $$ s^{2n} + c^{2n} = \left( \frac12 + u \right)^n + \left( \frac12 - u \right)^n. $$ Expanding using the binomial theorem, we find that all the odd powers of $x$ cancel and the even powers add up, so that we have an even polynomial in $x$. The degree of this polynomial is exactly $2 \lfloor n/2 \rfloor$.

At the next step, we replace $u^2$ by $\frac14 - s^2c^2$, so that we again get an even polynomial -- this time in $s \cdot c$ -- of the same degree, namely $2 \lfloor n/2 \rfloor$. In other words, we get a polynomial in $s^2 c^2$ of degree $\lfloor n / 2 \rfloor$. (However, the actual polynomials themselves do not seem very interesting.)

In general, one should not expect much more simplification. As Henning notes below, for $n =2$ and $n=3$, the expression $\lfloor n / 2 \rfloor$ is just $1$, which is why we got a polynomial of containing only a constant term and a $\sin^2 x \cos^2 x$ term. This pretty pattern disappears even for $n=4$.

share|improve this answer
2  
So the true pattern is that $\sin^{2n}+\cos^{2n}$ is a polynomial of degree $\lfloor n/2\rfloor$ in $\sin^2\cos^2$ with constant term $1$. (Odd powers of $x$ always cancel out after the binomial expansion). For $n=2$ and $n=3$ this polynomial is of degree $1$ with fixed constant term, so it looks like the beginning of a prettier pattern than it really is. –  Henning Makholm Dec 31 '11 at 15:10
    
Note: Thanks to a poor choice of notation, I had used $x$ in a different sense than the question; that has subsequently been changed to $u$ in my answer. The $x$ in @Henning's comment should be read as "$u$" for compatibility with my current answer. –  Srivatsan Dec 31 '11 at 16:02
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.