0

Artin states that:

Theorem 1.10: If $f(x)$ is log convex on a certain interval, and if $c$ is any real number $\neq 0$, then both the functions $f(x+c)$ and $f(cx)$ are log convex on the corresponding intervals.

I do not really understand the part of "on the corresponding intervals". Does it mean:

(Reformulation): If $f$ is log convex on $(a,b)$, and if $c\not = 0$, then $x\mapsto f(x+c)$ is convex on $(a-c,b-c)$ and $x\mapsto f(cx)$ is log convex on $(a/c,b/c)$?

By "log convex", he mean "logarithmically convex". A function $f:A\to \mathbb{R}$ is said to be logarithmically convex if $$ f(\lambda x+(1-\lambda)y)\leq f(x)^{\lambda}f(y)^{1-\lambda} $$ for all $x,y\in A$ and all $\lambda\in (0,1)$.

  • I think it is correct, up to (possibly) the order of $a/c$ and $b/c$ for $c <0$. – Aphelli Jan 30 '22 at 00:29
  • @Mindlack Thanks. What would happen, if $b=\infty$? Would it become: If $f$ is log convex on $(a,\infty)$, and if $c\not = 0$, then $x\mapsto f(x+c)$ is convex on $(a-c,\infty)$ and $x\mapsto f(cx)$ is log convex on $(a/c,\infty)$? (I assume $c>0$ to the last one, otherwise it would have been $(-\infty,a/c)$. Please correct me, if it's wrong) – Mr.MathDoctor Jan 30 '22 at 00:39
  • It should stay correct as well. – Aphelli Jan 30 '22 at 00:43

0 Answers0