I'm trying to find the exact solution to the equation $x+x^x=3$. I know the answer is approximately 1.4, but what is its formal definition? I understand it may not be from the result of a real function, since the exact solution to its brother, $x^x=3$, requires use of the Lambert W function for its formal definition ($x=e^{W(\ln 3)} $), which cannot be expressed in elementary terms. I have tried looking this equation up on Wolfram alpha but it can only approximate $x$. What would be the exact definition of this variable?
-
1What is the source of the problem and why numerical approximation is not satisfactory? I doubt the solution can be expressed using Lambert W. – Vasili Oct 04 '22 at 18:37
-
1im doing this more so out of interest and less related to practical applications – Jacob Oct 04 '22 at 18:42
-
1AFAIK, there's no "closed-form" expression. You'll just have to settle for the numerical solution $x \approx 1.3994607812259618$. – Dan Oct 04 '22 at 18:47
-
How did you arrive to this conclusion? – Jacob Oct 04 '22 at 18:52
-
2No, it is not, @K.defaoite That is the solution to $x^x=3.$ $x+x^x=3$ has a different solution. – Thomas Andrews Oct 04 '22 at 19:03
-
1In general, even with the $W$ function, you can't get closed formula for solutions of this kind of equation – Thomas Andrews Oct 04 '22 at 19:05
-
There are ways to approximate the solution. Since $x+x^x$ is increasing on $x>1,$ we can just do a binary search, for example, starting with the interval $[0,1].$ It might be faster to apply something like Newton's method. – Thomas Andrews Oct 04 '22 at 19:13
-
Even simple lagrange reversion fails here for inverting $x^x+x$ – Тyma Gaidash Oct 04 '22 at 19:14
-
No exact form :( – Max0815 Oct 04 '22 at 19:14
-
It converges slowly, but Lagrange reversion works here and here for inverting $x+\frac12 x^x$ – Тyma Gaidash Oct 04 '22 at 21:07
-
1@Tyma: That looks interesting! I didn't know about it, thanks! I usually only use basic Fixed Point Iteration, Newton's Method... – Matija Oct 04 '22 at 21:30
-
See also : https://tetrationforum.org/showthread.php?tid=1258 where $x^x = x $ is discussed. – mick Nov 22 '23 at 12:02
2 Answers
Irrational numbers are usually defined as solutions of equations, for example $x^2=2$, as limits, say $\sum_{k=0}^\infty(-1)^k\frac{4}{2k+1}$ or $\lim_{n\rightarrow\infty}(1+\frac{1}{n})^{1/n}$, or by introducing new notation, say $\sqrt{2},\pi$ or $e$. You stated the definition as a solution of an equation. Since introducing new notation is not helpful in this case, I only propose the following two alternatives. Let $x^*\in(0,\infty)$ be the solution of $x+x^x=3$.
- Limit: Consider the sequence given by $a_0=1$ and $a_{k+1}=\frac{3}{1+a_k^{a_k-1}}$ for $k\ge 0$. The sequence $(a_k)_k$ converges to the solution of the equation, i.e. $x^*=\lim_{k\rightarrow\infty}a_k$.
- Using the inverse function: Let $f:[1,\infty)\rightarrow(0,\infty)$, $x\mapsto x+x^x$, and notice that $f$ is strictly increasing since $f'(x)=1+(\ln(x)+1)x^x>2$, so the solution of the equation is $x^*=f^{-1}(3)$.
These would be three common ways (equation, limit, function value) to formally define the number $x^*$. I hope this helps!

- 3,526
$$x+x^x=3$$
We see, this equation is a polynomial equation of more than one algebraically independent monomials ($x,x^x$) and with no univariate factor. We therefore don't know how to rearrange the equation for $x$ by applying only finite numbers of elementary functions (elementary operations) we can read from the equation.
$$x+e^{x\ln(x)}=3$$ $$e^{x\ln(x)}=3-x$$ $$1=(3-x)e^{-x\ln(x)}$$ $$(3-x)e^{-x\ln(x)}=1$$
We see, we cannot solve the equation in terms of Lambert W because of the exponent of the exponential.
$x\to 3-t$: $$te^{(t-3)\ln(3-t)}=1$$
We see, the equation can be solved in terms of Hyper Lambert W.
$$G((t-3)\ln(3-t)e^{-t};t)=1$$ $$t=HW((t-3)\ln(3-t)e^{-t};1)$$ $$x=3-HW((t-3)\ln(3-t)e^{-t};1)$$
So we have a closed form for $x$, and the representations of Hyper Lambert W give some hints for calculating $x$.

- 6,964
-
-
Hyper Lambert W can be used i.a. for equations of the form $az+bf(z)=c$ or $(a+bz)f(z)=c$ if $f(z)$ is a tower of exponentials. – IV_ Oct 30 '23 at 15:45
-
See also : https://tetrationforum.org/showthread.php?tid=1258 where $x^x = x $ is discussed. – mick Nov 22 '23 at 12:02