0
$\begingroup$

Prove that for every positive integer $n$ and any real numbers $a_1,a_2,\cdots , a_n, b_1,\cdots, b_n $, the equation \begin{equation} \phantom{=}a_1 \sin(x)+a_2 \sin(2x)+\cdots +a_n \sin(nx) \\ = b_1 \cos(x)+b_2 \cos(2x)+\cdots +b_n \cos(nx)\end{equation} has at least one real solution.

I have tried to use the Intermediate Value Theorem and induction to solve this, but can't really prove it. My idea was to show that \begin{equation} g_n(x)=\sum_{k=1}^n a_k \sin(kx)-b_k\cos(kx) \end{equation} Is greater than zero at one point and less than zero at another one.
However, I got stuck at the inductive step, as it is not really clear to me whether \begin{equation} g_{n+1}(x)=a_{n+1}\sin([n+1]x)-b_{n+1}\cos([n+1]x)+g_n(x) \end{equation} Has the property that for some $x_1$ and $x_2$ in $\mathbb{R}$, $g_{n+1}(x_1)\geq 0$, and $g_{n+1}(x_2)\leq 0$ provided that $g_n(x)$ has this property.

I haven't taken any complex variable course yet, so I don't know if this would involve Fourier series or something related.

Another idea that I had was to integrate the function $g_n(x)$ from $0$ to $2\pi$ and show that it's zero. But I'm not sure if that works for every $a_i$ and $b_i$.

I would really appreciate any hint or sketch of proof.

Edits: I added my attempts of proving this.

$\endgroup$
4
  • 1
    $\begingroup$ $\int_0^{2\pi}g_n=\sum (a_k\int_0^{2\pi}\sin(kx)-b_k\int_0^{2\pi}\cos(kx))$. $\endgroup$ Commented Oct 13, 2020 at 2:46
  • 1
    $\begingroup$ You’re right about the integration idea; as Gerry has explicitly hinted above; (the idea, extended further, even gives you the bonus that there is a zero between $2\pi m$ and $2\pi(m+1)$ for each integer $m$.) $\endgroup$ Commented Oct 13, 2020 at 3:04
  • 1
    $\begingroup$ I see, this is an easier way of proving it than by using induction. Thank you very much! I really appreciate that you took the time to answer. $\endgroup$
    – P3p3O
    Commented Oct 13, 2020 at 3:07
  • $\begingroup$ Just using the fact that, for a continuous function $f$ in $[a,b]$, if $\int_a^bf(x)dx=0$, then $f(x)=0$ has at least a solution. $\endgroup$
    – xpaul
    Commented Oct 13, 2020 at 16:44

0

You must log in to answer this question.

Browse other questions tagged .