0
$\begingroup$

Let $B\subseteq \Bbb{R}^2$ be a closed ball of radius $\delta < 1$ centered at $(0,0)$. Let $f:B\to \Bbb{R}_{\geq 0}$ be real analytic and have only one zero, namely $(0,0)$. Moreover, assume that $f$ is strictly increasing along lines intersecting the origin.

Claim. Let $$ \tilde{f}(r,\theta) = f(r\cos(\theta), r\sin(\theta)) $$ be the polar form of $f$. Moreover, define $n_{\theta}\in \Bbb{N}$ to be the smallest non-negative integer such that $\tilde{f}^{(n_{\theta})}(0,\theta)\ne 0$. Lastly, let $\sup_{\theta\in [0,2\pi)}n_{\theta}\leq N <\infty$. Then, there exists some $r' \leq \delta$ and $\theta'\in[0,2\pi)$ such that for all $r\leq r'$,

$$ \tilde{f}(r,\theta) \geq \tilde{f}(r,\theta'). $$

Ideas. I'm thinking that maybe there's some sort of "cardioid" function where it decreases as it approaches $\theta = 0$. However, I'm unsure if such a function is real-analytic. Any advice would be appreciated.

$\endgroup$
8
  • 1
    $\begingroup$ I downvoted for now. What is the point of defining $n_\theta$ and $N$? Also $\theta^*$ should be $\theta'$ or vice-versa. $\endgroup$ Commented May 13 at 20:53
  • $\begingroup$ How about $f(x,y)=x^2+y^4$. I don't think this assumes its minimum on circles at one fixed $\theta^*$. $\endgroup$ Commented May 13 at 21:11
  • $\begingroup$ @mathworker21 The function I want to apply this claim to has that property. Does that property follow directly from the fact that $f$ is real analytic? If it does can you please enlighten me? It feels like you don't offer constructive criticism. $\endgroup$ Commented May 14 at 0:01
  • 1
    $\begingroup$ @Doofenshmert: I was sloppy with the details, but the general point (why would $r$ drop out of the minimization problem?) seems valid. How about $f=x^2+y^2+(y-x^2)^2$. The minimum on a circle is achieved at $y=x^2$, which is not at a constant angle. (We can also take a higher power of $y=x^2$ if there is a problem with the "increasing on radii" condition.) $\endgroup$ Commented May 14 at 0:48
  • 1
    $\begingroup$ @ChristianRemling Good counterexample. Thanks! $\endgroup$ Commented May 14 at 1:20

0

You must log in to answer this question.

Browse other questions tagged .