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 was told that the half-normal distribution, with density: $$ f(x)=\frac{1}{\sigma}\sqrt{\frac{2}{\pi}}\exp \left(-\frac{x^2}{2 \sigma^2} \right) \qquad \forall x \geq 0 $$ and the exponential distribution, with parameter $\lambda>0$ and density : $$ g(x) = \lambda \exp ( -\lambda x ) \qquad \forall x \geq 0 $$ cannot be parametrised (ie, for any $\lambda>0$, and $\sigma>0$) so that a constant $c > 0$ exists such that $g(x) \leq c f(x) \qquad \forall x \geq 0$

Is this true ? Is there a proof for this, especially one that I am likely to be able to understand (high school mathematics) ? If not, is there a counter-example ?

share|improve this question
add comment

1 Answer 1

up vote 1 down vote accepted

Since both $f(x)$ and $g(x)$ and non-negative for $x \geqslant 0$, the inequality $g(x) \leqslant c f(x)$ translates into $\log g(x) \leqslant \log(c) + \log f(x)$. That if such $c$ does exist, then there exists a $\kappa = \frac{c}{\lambda} \frac{1}{\sigma} \sqrt{\sqrt{2}{\pi}}$ such that $$ -\lambda x \leqslant -\frac{x^2}{2\sigma^2} + \log \kappa $$ That is $$ \frac{x^2}{2 \sigma^2 } - \lambda x = \frac{1}{2 \sigma^2} \left(x- \sigma^2 \lambda\right)^2 - \frac{\lambda^2 \sigma^2}{2} \leqslant \log \kappa $$ It is now clear such $\kappa$ can not exist, since $(x-\sigma^2 \lambda )^2$ is unbounded.

share|improve this answer
add comment

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.