2
$\begingroup$

The present discussion arises from this MO question. Below, $e$ stands for Euler's number and let $$\tau:=\arccos\left(\frac{\sin e-\sin 1}{e-1}\right)\approx 1.82\cdots.$$

An application of the Mean Value Theorem (for derivatives) to the function $f(t)=\sin t$ leads to $$\frac{\sin(e\,t)-\sin(t)}{e\,t-t}=\cos(\xi_tt) \qquad \text{for some $1\leq\xi_t\leq e$}. \tag1$$

QUESTION. Is it true that for each $t>0$, one can always find some $\xi_t\geq\tau$ such that (1) holds? Example: $\xi_1=\tau$.

$\endgroup$
2
  • 1
    $\begingroup$ You found such $\xi_t\in[1,e]$ in (1). Now add to it $2\pi N/t$ where $N$ is large, and you obtain new $\xi_t>\tau$ satisfying (1). $\endgroup$ Mar 12, 2020 at 11:25
  • $\begingroup$ Given that the answers (1 2) to your question indicate that it's not what you meant, you should probably change it so that it says what you mean. $\endgroup$
    – LSpice
    Mar 12, 2020 at 20:07

2 Answers 2

5
$\begingroup$

In view of the answer by Carlo Beenakker and the comment by Alexandre Eremenko, it appears that what you actually had in mind is the following question:

By the mean value theorem, for each $t\in(0,1]$ there is some $\xi_t\in(1,e)$ such that \begin{equation*} r(t):=\frac{\sin et-\sin t}{(e-1)t}=\cos(\xi_t t). \tag{2} \end{equation*} (Since $\cos u$ is strictly decreasing in $u\in[0,e]$, the value of $\xi_t$ is unique for each $t\in(0,1]$.) Is it true that $\xi_t\ge\tau$ for all $t\in(0,1]$?

The answer to this question is yes. Indeed, for $t\in(0,1)$ we have $\xi_t t\in(0,e)\subset[0,\pi]$ and $\tau t\in(0,\tau]\subset[0,\pi]$. Therefore, in view of (2) and because $\cos$ is strictly decreasing on $[0,\pi]$, we see that \begin{equation*} \xi_t>\tau\iff d(t):=\cos\tau t-r(t)>0; \tag{3} \end{equation*} here and in what follows, $t\in(0,1)$. Next, \begin{equation*} d_1(t):=(e-1) d(t)/t^2=\sum_{j=1}^\infty(-1)^jb_j t^{2j-2}-\sum_{j=1}^\infty(-1)^ja_j t^{2j-2}, \end{equation*} where \begin{equation*} a_j:=\frac{e^{2 j+1}-1}{(2 j+1)!},\quad b_j:=\frac{(e-1) \tau^{2 j}}{(2 j)!}. \end{equation*} It is easy to see that $0<a_j<a_{j+1}$ and $0<b_j<b_{j+1}$ for all natural $j$. So, \begin{equation*} d_1(t)>-b_1+b_2t^2-b_3t^4+a_1-a_2t^2>0 \quad\text{if}\quad 0<t\le4/5. \end{equation*}

It remains to prove that \begin{equation} d_2(t):=(e-1)t d(t)>0\quad\text{if}\quad 4/5<t<1. \end{equation} Since $d_2(1)=0$, it suffices to show that \begin{equation} d_2'(t)=\cos t-e \cos et+(e-1) \cos \tau t-(e-1) \tau t \sin \tau t<0 \end{equation} for $t\in(4/5,1)$. Since $\cos t,\cos et,\cos \tau t$ are decreasing in $t\in(4/5,1)$ and $\sin \tau t$ is concave in $t\in(4/5,1)$, the desired result follows because for $t\in[t_j,t_{j+1}]$ and $j=0,\dots,n-1$ \begin{equation} d_2'(t)\le\cos t_j-e \cos et_{j+1}+(e-1) \cos \tau t_j -(e-1) \tau t_j \min(\sin \tau t_j,\sin \tau t_{j+1})<0, \end{equation} where $n:=20$ and $t_j:=4/5+j/(5n)$.


To illustrate the above proof, here is the graph $\{(t,d(t))\colon0<t<1\}$:

enter image description here

$\endgroup$
1
  • $\begingroup$ This is indeed what I meant, although I missed specifying the range $0<t<1$. Cool job! $\endgroup$ Mar 12, 2020 at 17:59
4
$\begingroup$

this is a plot of $\frac{\sin e\,t-\sin t}{e\,t-t}-\cos\xi \,t$ as a function of $\xi$ for $t=1.1$; the curve does not cross zero in the interval $[\tau,e]$, so I would conclude that (1) does not hold.

$\endgroup$
1
  • $\begingroup$ This is true due to my mistake in not saying $0<t<1$. Thanks. $\endgroup$ Mar 12, 2020 at 17:54

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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