Find all continuous functions $f\colon \mathbb{R}_{>0}\to \mathbb{R}_{\geq 1}$ for which the following equation holds for all positive reals $x$, $y$: \[f\left(\frac{f(x)}{y}\right)-f\left(\frac{f(y)}{x}\right)=xy\left(f(x+1)-f(y+1)\right)\]
Problem
Source: 2024 Israel TST Test 3 P3
Tags: functional equation, algebra, continuous function, Positive reals, function
01.02.2024 14:48
My problem! Created it with a friend.
01.02.2024 17:36
Phorphyrion wrote: Find all continuous functions $f\colon \mathbb{R}_{>0}\to \mathbb{R}_{\geq 1}$ for which the following equation holds for all positive reals $x$, $y$: \[f\left(\frac{f(x)}{y}\right)-f\left(\frac{f(y)}{x}\right)=xy\left(f(x+1)-f(y+1)\right)\] Can you send full test =((
01.02.2024 18:05
Quote: Case 1: $f$ is bounded in $[1,\infty]$ Substitute $y=\frac{f(x)}{z}$ and multiply by $z$ to get: $$zf(z)-zf \left (\frac{f(\frac{f(x)}{z})} {x}\right) = xf(x)\left((f(x+1)-f\left(\frac{f(x)}{z}+1\right)\right)$$We can take $z \to 0$ to get: $$zf(z)-O(1) = xf(x)\left((f(x+1)-O(1)\right) \implies zf(z) = O(1)$$ Why is $f \left (\frac{f(\frac{f(x)}{z})} {x}\right) \sim O\left(\frac{1}{z}\right)$
01.02.2024 18:18
tadpoleloop wrote: Quote: Case 1: $f$ is bounded in $[1,\infty]$ Substitute $y=\frac{f(x)}{z}$ and multiply by $z$ to get: $$zf(z)-zf \left (\frac{f(\frac{f(x)}{z})} {x}\right) = xf(x)\left((f(x+1)-f\left(\frac{f(x)}{z}+1\right)\right)$$We can take $z \to 0$ to get: $$zf(z)-O(1) = xf(x)\left((f(x+1)-O(1)\right) \implies zf(z) = O(1)$$ Why is $f \left (\frac{f(\frac{f(x)}{z})} {x}\right) \sim O\left(\frac{1}{z}\right)$ Assume that $f(x)$ is bounded by $M$ in $[1,\infty]$, then $M\geq f(\frac{f(x)}{z}) \geq 1$, which means that $ \frac{M}{x} \geq \frac{f(\frac{f(x)}{z})} {x} \geq \frac{1}{x}$. Because a continuous function is bounded in a closed segment, it means that $f \left (\frac{f(\frac{f(x)}{z})} {x}\right)=O(1)$, and in particular $f \left (\frac{f(\frac{f(x)}{z})} {x}\right)=O(\frac{1}{z})$
02.02.2024 00:25
Quote: Assume for contradiction that $\frac{f(x_0)}{y}<\frac{f(y)}{x_0}$ for all small enough $y$. In particular. $\frac{f(y)}{x_0} \to \infty$ when $y \to 0$, and because $f$ is continuous and unbounded in $[1,\infty]$ we can take $\frac{f(y)}{x_0}$ to be a global maximum of $f$ in the segment $[1,\frac{f(y)}{x_0}]$, which means that $$f\left(\frac{f(x_0)}{y}\right)<f\left(\frac{f(y)}{x_0}\right)$$Which is a contradiction. Therefore, $\frac{f(x_0)}{y}>\frac{f(y)}{x_0}$ for all small enough $y$, which proves the lemma. I am feeling a bit skeptical by the argument here. Speaking of the limits in very loose terms with several variables in motion. Could you be more explicit? Edit:
02.02.2024 00:58
tadpoleloop wrote: Quote: Assume for contradiction that $\frac{f(x_0)}{y}<\frac{f(y)}{x_0}$ for all small enough $y$. In particular. $\frac{f(y)}{x_0} \to \infty$ when $y \to 0$, and because $f$ is continuous and unbounded in $[1,\infty]$ we can take $\frac{f(y)}{x_0}$ to be a global maximum of $f$ in the segment $[1,\frac{f(y)}{x_0}]$, which means that $$f\left(\frac{f(x_0)}{y}\right)<f\left(\frac{f(y)}{x_0}\right)$$Which is a contradiction. Therefore, $\frac{f(x_0)}{y}>\frac{f(y)}{x_0}$ for all small enough $y$, which proves the lemma. I am feeling a bit skeptical by the argument here. Speaking of the limits in very loose terms with several variables in motion. Could you be more explicit? Edit:
Yeah, your explanation works, you just need to say that for small enough $t$, $\frac{f(x_0)}{t}$ is in the range in which $f(\frac{f(t)}{x_0})$ is a global maximum, and you can take such $t$, because there are such maximums as big as you want
02.02.2024 05:38
Zsigmondy wrote: My problem! Created it with a friend.
I don't understand the argument here , you seem to prove later in your proof that $lim_{x\to \infty} f(x)=\infty$ but then in you first part taking $z\to 0$ and $x$ fixed you shoud have a problem in the term $f(\frac{f(x)}{z}+1)$ how com you consider it $O(1)$ ? Maybe I am missing something.
02.02.2024 11:10
oty wrote: Zsigmondy wrote: My problem! Created it with a friend.
I don't understand the argument here , you seem to prove later in your proof that $lim_{x\to \infty} f(x)=\infty$ but then in you first part taking $z\to 0$ and $x$ fixed you shoud have a problem in the term $f(\frac{f(x)}{z}+1)$ how com you consider it $O(1)$ ? Maybe I am missing something. I assume in that part that $f$ is bounded in $[1,\infty]$. I also never proved that $lim_{x\to \infty} f(x)=\infty$, in the proof of the first lemma I divide into cases based on whether $f$ is bounded on $[1,\infty]$ and later I divide into cases based on whether $f$ has a limit in $\infty$ (and in fact I prove that there are no solutions when $lim_{x\to \infty} f(x)=\infty$).