Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Solving for a*x=b-c*log(d*x)?

  1. Mar 22, 2014 #1

    Uan

    User Avatar

    Not really sure where this question belongs in this forum...

    I was solving an engineering problem and I got to the form

    [tex]ax=b-cln(dx)[/tex]

    where a, b, c and d are constant real values. I had a peek at the answer and they got a unique positive real valued answer for x but I have no idea how. Some searching I came across the Lambert W-Function and I got it into the form

    [tex]\frac{1}{d}e^{\frac{b}{c}} = xe^{\frac{ax}{c}} [/tex]

    How do I proceed to apply the Lambert W-Function from here?

    WolframAlpha found that

    [tex]x = \frac{c}{a}W\left ( \frac{a}{cd}e^{\frac{b}{c}} \right )[/tex]
     
  2. jcsd
  3. Mar 22, 2014 #2

    mfb

    User Avatar
    Insights Author
    2015 Award

    Staff: Mentor

    $$x=W(g) \Leftrightarrow x*e^x = g$$
    Therefore,
    $$\frac{ax}{c} = W\left ( \frac{a}{cd}e^{\frac{b}{c}} \right ) \Leftrightarrow \frac{ax}{c} \exp\left(\frac{ax}{c}\right) = \frac{a}{cd}e^{\frac{b}{c}}$$

    Multiply both sides with c/a and you get the equation in your post. The other direction works the same, you just have to see that you need the shape (something)*e^(something) and work with the constants to get that.
     
  4. Mar 22, 2014 #3

    Shayan.J

    User Avatar
    Insights Author
    Gold Member

    Well, LambertW is defined to be the inverse of [itex] W e^W [/itex]. If there was a way to invert that function, there would be no need to define LambertW! The expression you got is the most thing you can do. Then you are supposed to say "Oh yeah, so x should be the LambertW of such and so". And for actually finding numbers for x, you should consult mathematical tables or math softwares .
     
  5. Mar 22, 2014 #4

    Uan

    User Avatar

    Thanks, that clears a lot up. So really the Lambert W in this case doesn't help all that much - just allows the function to be in a more recognisable form. You still need to go back to...

    [tex]\frac{1}{d}e^{\frac{b}{c}} = xe^{\frac{ax}{c}} [/tex]

    and solve numerically for x.
     
  6. Mar 23, 2014 #5

    mfb

    User Avatar
    Insights Author
    2015 Award

    Staff: Mentor

    Or look up function values of the Lambert W function, yes.
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?
Draft saved Draft deleted



Similar Discussions: Solving for a*x=b-c*log(d*x)?
  1. Solve x^3*e^(-a/x)=b (Replies: 5)

  2. Y=mx+b Solving for m (Replies: 7)

Loading...