1
$\begingroup$

I have been working on this problem that I have heard was posted in a high-school pre-calculus course. As I would expect, there should be a closed-form expression for $e^x + ln(x) = 4$ for some $x$, but WolframAlpha does not give one, only an approximate numerical answer.

Is there a way of solving this without the use of calculus or any other higher-order methods?

I have been able to get it to $xe^{e^x} = e^4$, but that does not seem to help anything.

$\endgroup$
  • $\begingroup$ I think this one is out of the reach of even the Lambert W function, which does provide "closed forms" for things like $\ln x = 1/x.$ $\endgroup$ – coffeemath Nov 4 '14 at 2:51
  • 3
    $\begingroup$ You aren't going to find a nice closed form solution. $\endgroup$ – Thomas Andrews Nov 4 '14 at 2:51
  • $\begingroup$ Even Lambert won't help you with this one, I bet. @coffeemath $\endgroup$ – Thomas Andrews Nov 4 '14 at 2:51
  • $\begingroup$ If $e^x$ is near $4$, then $x\ge 1$ and $\ln x\ge 0$, so we know that both terms supply a positive contribution on the LHS. Suppose we try $1\le x\le \ln 4=2\ln 2\le 2$... Is there a series of approximations that gets us close? $\endgroup$ – abiessu Nov 4 '14 at 2:54
  • $\begingroup$ @ThomasAndrews Yes the OP equation won't get a W solution. However the simpler $\ln x=1/x$ does have one, as I recall. A fellow grad student once wrote it on the board, having forgotten to include $d/dx$ in the statement $d/dx \ln x=1/x,$ and we went to work on it, and found some solution using Lambert. [That was years ago, I may have misremembered it.] $\endgroup$ – coffeemath Nov 4 '14 at 3:28
1
$\begingroup$

If the problem was to find the solution of $$e^x+x=4$$ the solution would have been $$x=4-W\left(e^4\right)$$

If the problem was to find the solution of $$x+\log(x)=4$$ the solution would have been $$x=W\left(e^4\right)$$

But the problem is to find the solution of $$e^x + ln(x) = 4$$ for which it does not seem to have an analytical solution. However, solving numerically shows that the solution is $x\approx 1.31531$ and looking at $RIES$, this number is very close to $$x_0=\frac{\sqrt{e+\frac{1}{\sqrt{e}}}}{ \log (4)} $$

To polish the solution, let us use one single iteration of Newton and get $$x_1=\frac{\sqrt{\frac{1}{\sqrt{e}}+e}}{\log (4)}-\frac{-4+e^{\frac{\sqrt{\frac{1}{\sqrt{e}}+e}}{\log (4)}}+\frac{1}{2} \log \left(\frac{1}{\sqrt{e}}+e\right)-\log (\log (4))}{e^{\frac{\sqrt{\frac{1}{\sqrt{e}}+e}}{\log (4)}}+\frac{\log (4)}{\sqrt{\frac{1}{\sqrt{e}}+e}}}$$

It should be notice that $f(x_1)=3.711\times 10^{-11}$ while $f(x_0)=-2.179\times 10^{-5}$

Not the exact solution but not too far !

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.