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 have a function $f(x_1,x_2) \colon \mathbb{R}^2_{+} \to \mathbb{R}_{+}$ positive homogenous: $$ f(\lambda x_1, \lambda x_2) = \lambda f(x_1,x_2), \; \forall \lambda > 0 $$ and such that $f(x_1,x_2)$ permits decomposition $$ f(x_1,x_2) = h(g(x_1)+g(x_2)) $$ where $h,g$ are some continuous functions. One of appropriate functions is $$ f_{0}(x_1,x_2) = C(x_1^{\gamma}+x_{2}^{\gamma})^{\frac{1}{\gamma}}, \; \gamma > 0, \; C \geq 0 $$ Are there some other functions that satisfy specified conditions or $f_{0}(x_1,x_2)$ is unique?

Update If $h = g^{-1}$ then $f_{0}$ is a unique family of solutions (Hardy, G. H.; Littlewood, J. E.; Pólya, G. (1952) Inequalities. 2nd ed, page 68).

share|improve this question
Well, $f_0$ isn't unique, because you can choose $C$ and $\gamma$. –  Stefan Nov 21 '12 at 10:03

1 Answer 1

The condition $f( \lambda x_1, \lambda x_2) = \lambda f(x_1, x_2)$ for $\lambda > 0$ for $x_1, x_2 > 0$ means $f$ is uniquely determined by its restriction to $S$, the part of the unit circle in the first quadrant, i.e. $\{(\cos \theta, \sin \theta): \theta \in (0, \frac{\pi}{2}) \}$. Working in polar coordinates, it is easy to see that $f$ is smooth iff it's restriction to $S$ is.

With that, on $S$, the second condition is $f = h(g(\cos \theta) + g(\sin \theta))$, for any choice of continuous $h,g$.

It suffices to find such an $f$ which is not smooth. Indeed, one should be able to do it by taking $g = id$, note the map $(0, \frac{\pi}{2}) \rightarrow \mathbb{R}$ given by $\theta \mapsto \sin \theta + \cos \theta$ is a local diffeomorphism for $\theta \neq \frac{\pi}{4}$, so it suffices to take $h$ to be any continuous function $\mathbb{R} \rightarrow \mathbb{R}$ which isn't differentiable at say the image of $\frac{\pi}{8}$.

share|improve this answer
Note that $g$ cannot be $x\mapsto ax$ and $x \mapsto bx$ at the same time, unless $a = b$, which makes your $f$ equal to $Ca\bigl(x_1^\gamma + x_2^\gamma)^{1/\gamma}$. So, nothing new. –  martini Nov 21 '12 at 9:39

Your Answer


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.