10

This equation clearly cannot be solved using logarithms.

$$3 + x = 2 (1.01^x)$$

Now it can be solved using a graphing calculator or a computer and the answer is $x = -1.0202$ and $x=568.2993$.

But is there any way to solve it algebraically/algorithmically?

Later
  • 722
  • 2
  • 8
  • 24
  • 2
    I'm not sure why this question was downvoted. It seems like a legitimate question to me. – EuYu Oct 20 '12 at 05:32
  • Are you familiar with the Lambert-W Function (http://mathworld.wolfram.com/LambertW-Function.html), hence see: http://www.wolframalpha.com/input/?i=Solve%5B3+%2B+x+%3D+2+%281.01+%5E+x%29%2Cx%5D – Amzoti Oct 20 '12 at 05:37
  • (I didn't downvote) maybe because there are similar problems already on here? I swear I've seen a few, but I just searched & turned up nothing, so I can't blame you for not finding one! Similar one: http://math.stackexchange.com/questions/61774/how-do-i-solve-this-equation-involving-a-logarithm –  Oct 20 '12 at 05:39

3 Answers3

15

I have solved a question similar to this before. In general, you can have a solution of the equation

$$ a^x=bx+c $$

in terms of the Lambert W-function

$$ -\frac{1}{\ln(a)}W_k \left( -\frac{1}{b}\ln(a) {{\rm e}^{-{\frac {c\ln(a) }{b}}}} \right)-{\frac {c}{b}} \,.$$

Substituting $ a=1.01 \,,b=\frac{1}{2}\,,c=\frac{3}{2}$ and considering the values $k=0$ and $k=-1$, we get the zeroes $$x_1= -1.020199952\,, x_2=568.2993002 \,. $$

5

Polynomials don't play nice with exponentials, so no. If you work hard, you might find an answer in terms of the Lambert W function, but if I did I wouldn't feel much more enlightened.

Ross Millikan
  • 374,822
2

A standard root finding procedure (such as Newton's method) should solve the problem for you. You might also be interested in the Lambert W function, which will give you a "closed form" solution, assuming you have access to that function of course.

EuYu
  • 41,421