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

Exponential equation-Lambert function?

  1. Mar 2, 2007 #1
    Hallo

    I am trying to solve for x the equation

    (b-x)exp^[(a-x)]+c-x=0

    where a,b and c are constants.

    Could anyone tell me if this equation can be solved using the Lambert function?
    Thanks
     
  2. jcsd
  3. Mar 2, 2007 #2

    uart

    User Avatar
    Science Advisor

    Can the solution to that equation be expressed in terms of the Labert W function?

    Yes.
     
  4. Mar 2, 2007 #3

    uart

    User Avatar
    Science Advisor

    Hmmm. At first glance I thought yes, but now I'm not so sure.

    Certianly anything that can be expressed (with suitable change of variable) as a linear combination of log(x), x and 1 or as a linear combination of exp(x), x and 1, can be solved pretty easily in terms of W. I'm not 100% sure about linear combinations of x*exp(x), exp(x), x and 1 however.
     
  5. Jul 9, 2007 #4
    I don't know if the solution is accurate or not but here is the method.

    I have used a couple of approximations and assumptions.

    Assuming a > 0 and for x << a

    we can rewrite the equation as

    exp(a-x) = (x-c)/(b-x) ;

    adding one two both sides :

    1 + exp(a-x) = (b-c)/(b-x);

    and if a>> x then we can neglect 1 w.rt. exp(a-x) and so,

    (b-x)exp(a-x) = b - c;

    which yields x = b - W((b-c)exp(b-a));


    Can someone check this though? Also, does neone know if we can solve

    y = a*exp(b(x - y)) + c*exp(d(x-y));

    in terms of W?
     
  6. Jul 12, 2007 #5
    Hi,

    I have a simpler but related question.

    How do we solve the following equation:

    exp(ax)=bx+c

    Thanks.
     
  7. Jul 12, 2007 #6
    exp(ax) = bx + c;

    I think this might be the solution:

    let bx + c = y/a;

    a*x = (y - c*a)/b;

    exp(y/b)*exp(-c*a/b) = y/a;

    or

    -(a/b)*exp(-c*a/b) = (-y/b)*exp(-y/b);

    or -y/b = W(-(a/b)*exp(-c*a/b));

    or y = -b*W(-(a/b)*exp(-c*a/b));

    where y = a*(bx+c);

    You can check for mistakes
     
  8. Jul 12, 2007 #7
    Thanks a lot. Romain
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?



Similar Discussions: Exponential equation-Lambert function?
  1. Lambert W Function (Replies: 14)

  2. Lambert W function (Replies: 2)

Loading...