MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

$x\ln (\frac{2}{x})=k(a-x) \ln (\frac{2}{a-x})$

where $a$ and $k$ are positive constants. $a$ is usually small, say, $0< a<0.1$ and $x\in (0,a)$.

There are ways to calculate numerical solutions. However, as it is a intermediate step of an optimization problem, I do need closed form results to move forward with some strict proof. Does that exist?

One observation is that function $f(x)=x\ln (\frac{2}{x})$ is increasing when $x\in (0,2/e)$, thus with $0< a<0.1$ and $x\in (0,a)$, $f(x)$ is increasing. I don't know if this helps.

share|cite|improve this question
No chance unless there is some arcane special function you can write it in terms of. – Brendan McKay Apr 4 '12 at 7:39
@ Brendan - Are you teasing the OP?! Looks like homework to me. Differentiate twice and you've got the solution no? – Kevin Smith Apr 4 '12 at 10:11
Generally, the solution to the derivative of an equation is not a solution to the equation itself. – Gerald Edgar Apr 4 '12 at 12:52

Rather than solving for $x$ in terms of $a$, how about solving for $a$ in terms of $x$ like this: $$ a = x + \frac{x \operatorname{ln}(x/2)}{k \mathrm{W}\left(\frac{x \operatorname{ln} (x/2)}{2 k}\right)} $$ where $\mathrm{W}$ is the Lambert W function.
[Inspired by the solution to What is the name of $\frac{e^z-1}{z}$ and how to invert it? ]

share|cite|improve this answer

This isn't a research level mathematics question, but you do have to do a bit of high school calculus. Since $y\log(2/y)$ is strictly increasing and $k(a-y)\log(2/(a-y))$ is strictly decreasing on $(0,a)$, dividing the l.h.s by the r.h.s, differentiating twice and setting the result to $0$ should give you the unique point where the curves meet.

EDIT: This is incorrect (or at least it only holds for $k=1$, which trivially gives you the mid-point) because the inflection point is unchanged by the constant $k$. Apologies.

Speaking of arcane special functions! the $q$-digamma function function might be of some use:

$$\frac{d}{dz} \log\Gamma_q(z)=-\log(1-q)+\log(q)\sum_{0}^{\infty}\frac{q^{n+z}}{1-q^{n+z}}.$$

share|cite|improve this answer
Why is this question community wiki? – Kevin Smith Apr 4 '12 at 16:11
Questions automatically become CW if they have more than a certain number of revisions. Details are available in the site documentation somewhere. I don't know whether that feature accounts for what happened here. – Gerry Myerson Apr 5 '12 at 4:45

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.