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.

I'm looking for the derivative $\frac{d}{dx}\lceil 1/x \rceil$.

I would like to find a real number $1<x \le y$ satisfying the minimum of $\left\lceil \frac{y}{x} \right\rceil x$, when $y$ is a fixed value $>0$.

Is the ceiling function a problem ?

share|improve this question
    
Thank you. $x$ can be taken $\ge 1$. If this ceiling function is a problem, is there another method to determine a minimum ? –  Dingo13 Oct 14 '13 at 22:41
    
Sorry I've forgotten a term in my problem. Thank you again for the help. –  Dingo13 Oct 14 '13 at 22:43
1  
Ah, that makes a huge difference. I would drop your first and last sentences and just state the real problem: derivative won’t help. –  Brian M. Scott Oct 14 '13 at 22:44
    
The problem was indeed easy without this missing term. –  Dingo13 Oct 14 '13 at 22:45
    
Maybe there is a method to approximate correctly for the argument of minimum, for instance by bounding the ceiling function. No ? If for instance I look the derivative for the upper bound ? –  Dingo13 Oct 14 '13 at 22:48

2 Answers 2

Let $f(x)=\left\lceil\frac{y}x\right\rceil x$. If $x>y$, then $f(x)=x$, so $f(x)$ can be made arbitrarily large.

If we limit $x$ to the interval $[1,y]$, however, things are more interesting. Let $n=\lceil y\rceil$; as $x$ runs from $1$ to $y$, $\left\lceil\frac{y}x\right\rceil$ runs from $n$ down to $1$. Let $x_n=1$, and for $k=1,2,\ldots,n-1$ let $x_k=\frac{y}k$; then

$$\left\{x\in[1,y]:\left\lceil\frac{y}x\right\rceil=k\right\}=[x_k,x_{k-1})\;.$$

For $2\le k\le n$ the function $f(x)$ is increasing on $[x_k,x_{k-1})$, and

$$\lim_{x\to x_{k-1}^-}f(x)=kx_{k-1}=\frac{ky}{k-1}\;,$$

though this limit is not attained. Finally, $\frac{k}{k-1}$ is maximized at $k=2$, so as $x$ approaches $x_1=y$ from the left, $f(x)$ approaches $2y$ from below. The function $f(x)$ does not actually attain a maximum on $[1,y]$, but it is bounded above by $2y$ and attains values arbitrarily close to this.

Added: Somehow I managed to read minimum as maximum. Of course the analysis above shows that the minimum on $[x_k,x_{k-1})$ is $$f(x_k)=k\frac{y}{x_k}=y\;,$$ so the function achieves a minimum value of $y$ at each $x_k$.

share|improve this answer
    
Thank you, but in fact I'm looking for a minimum, not a maximum. –  Dingo13 Oct 15 '13 at 9:06
    
I'm looking for the minimum when $1<x\le y$. –  Dingo13 Oct 15 '13 at 9:50
    
@Dingo13: That information was already available in what I’d written, but I’ve made it explicit and corrected some errors resulting from trying to type on a Kindle. –  Brian M. Scott Oct 15 '13 at 10:15

$\large x > 0:$

\begin{align} \left\lceil x\right\rceil &= \Theta\left(x\right)\Theta\left(1 - x\right) + 2\Theta\left(x - 1\right)\Theta\left(2 - x\right) + 3\Theta\left(x - 2\right)\Theta\left(3 - x\right) + \cdots \\[3mm]&= \sum_{n = 0}^{\infty}\left(n + 1\right)\Theta\left(x - n\right) \Theta\left(n + 1 - x\right) \end{align}
\begin{align} {{\rm d}\left\lceil x\right\rceil \over {\rm d}x} &= \sum_{n = 0}^{\infty}\left(n + 1\right)\left[% \delta\left(x - n\right) \Theta\left(n + 1 - x\right) - \Theta\left(x - n\right) \delta\left(n + 1 - x\right) \right] \\[3mm]&= \sum_{n = 0}^{\infty}\left(n + 1\right)\left[% \delta\left(x - n\right) \Theta\left(1\right) - \Theta\left(1\right) \delta\left(n + 1 - x\right) \right] \\[3mm]&= \sum_{n = 0}^{\infty}\left(n + 1\right) \delta\left(x - n\right) - \sum_{n = 1}^{\infty} n\,\delta\left(n - x\right) = \sum_{n = 0}^{\infty}\delta\left(x - n\right) \end{align}
\begin{align} {{\rm d}\left\lceil x\right\rceil \over {\rm d}x} &= \sum_{n = 0}^{\infty}\delta\left(x - n\right)\,, \qquad \left\lceil x \right\rceil = 1\ \mbox{when}\ 0 < x < 1 \end{align}
\begin{align} {{\rm d}\left\lceil 1/x\right\rceil \over {\rm d}x} &= \sum_{n = 0}^{\infty}\delta\left({1 \over x} - n\right)\, \left(-\,{1 \over x^{2}}\right) = \sum_{n = 1}^{\infty} {\delta\left(x - 1/n\right) \over \left\vert -1/x^{2}\right\vert}\, \left(-\,{1 \over x^{2}}\right) \end{align}
\begin{align} \color{#ff0000}{\large{{\rm d}\left\lceil 1/x\right\rceil \over {\rm d}x}} &\color{#000000}{\large\ =\ } \color{#ff0000}{\large -\sum_{n = 1}^{\infty}\delta\left(x - {1 \over n}\right)} \end{align}
share|improve this answer
    
Thank you for the derivative. I suppose that $\delta$ is the Dirac function which is valued at 1 for 0 and 0 otherwise. As Brian has said, this derivative does not help a lot. –  Dingo13 Oct 15 '13 at 9:08

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.