$$x^{\cfrac{1}{x}}=(1+x)^{\cfrac{1}{1+x}}$$
Domain: $(0,+\infty)$
I know a numerical solution of $x\approx 2.293$
Does it have any analytical solutions?
If not, is it possible to prove that it doesn't have any analytical solutions?

- 184
-
You probably want to restrict your domain to $(0,\infty)$, so that both sides of the equation are well-defined real numbers. – diracdeltafunk Mar 13 '23 at 01:16
-
@diracdeltafunk thanks – SundayLi Mar 13 '23 at 01:18
-
I'm quite sure the answer is no, there is no way to express this solution in terms of elementary functions and constants. I don't know how to prove this, though. – diracdeltafunk Mar 13 '23 at 01:21
-
A systematic reflex in this case should be to take the logarithm on both sides... – Jean Marie Mar 13 '23 at 11:25
-
ContourPlot3D[ z == x^(y/x) - (y + x)^(y/(y + x)), {y, -2, 1}, {x, 1, 3}, {z, -0.1, 0}] reveals no other root when (y,z) domains are extended – Narasimham Mar 18 '23 at 21:14
2 Answers
If $\frac{log(x)}{x}= \frac{log(1+x)}{1+x}$ then $log(x)+xlog(x)=xlog(1+x)$ so $log(x) + x(log(x)-log(x+1)) = 0$ so $log(x) + xlog(\frac{x}{x+1}) = 0$. so $log(x) + log((\frac{x}{x+1})^x) = 0$ so $log(x(\frac{x}{x+1})^x) = 0$ so $(\frac{x}{x+1})^x = \frac{1}{x}$ so $$x^{x+1}= (x+1)^x$$ aswell. So like maybe another way to state it is that its the fixed point of the mapping $(1+\frac{1}{x})^x$ So like its well known that this limit approaches e, but idk about like fixed points. But hopefully this is helpful somewhat.
You might also notice that this is a contraction mapping around your point. So if you want to approximate it very quickly, recursively apply the function $f(x)= (1+\frac{1}{x})^x$ to itself over and over starting at x=2 (this is a result of banach fixed point thm). So like its equal to $$\lim_{i\to \infty} f^i(2)$$

- 347
-
This also works if you instead iterate the function on x=1, and if you try to write out the iterations, this takes care of like the exponent, but its like some weird continued fraction. – Allan Henriques Mar 13 '23 at 03:22
-
Thanks. I just started from the iteration and reached the question I asked. – SundayLi Mar 13 '23 at 03:51
I'd bet money that there is no way to express this solution in terms of elementary functions and constants. I don't know how to prove this.
However, I can prove that the solution is unique, which might be useful to you.
Note that $x$ is a solution to the equation iff $(x+1)\log(x) - x\log(x+1) = 0$. But now notice that $$\frac{\mathrm{d}}{\mathrm{d}x} ((x+1)\log(x) - x\log(x+1)) = \frac{2 x+1}{x^2+x}+\log (x)-\log (x+1)$$ so $$\frac{\mathrm{d}^2}{\mathrm{d}x^2} ((x+1)\log(x) - x\log(x+1)) = -\frac{x^2+x+1}{x^2 (x+1)^2}.$$ When $x > 0$, $-\frac{x^2+x+1}{x^2 (x+1)^2} < 0$, so $\frac{\mathrm{d}}{\mathrm{d}x} ((x+1)\log(x) - x\log(x+1))$ is a decreasing function. But also $$\lim_{x \to \infty} \frac{\mathrm{d}}{\mathrm{d}x} ((x+1)\log(x) - x\log(x+1)) = \lim_{x \to \infty} \left(\frac{2 x+1}{x^2+x}+\log (x)-\log (x+1)\right) = 0,$$ so we conclude that $\frac{\mathrm{d}}{\mathrm{d}x} ((x+1)\log(x) - x\log(x+1))$ is always positive. Thus, $(x+1)\log(x) - x\log(x+1)$ is an increasing function, so it has at most one zero.

- 14,911