How to Solve the Integral of e^x ln(x) dx?

  • Context: Undergrad 
  • Thread starter Thread starter ultrazyn
  • Start date Start date
  • Tags Tags
    Integral
Join the discussion
Registration is free. Ask a follow-up in this thread, or start your own.
9 replies · 3K views
ultrazyn
Messages
2
Reaction score
0
i need help with this integral

[tex]\int{e^x}{ln{x}}dx[/tex]

when i let u=ln(x) and v=e^x, i get

[tex]\{e^x}{lnx}\ - \int{e^x/{x}}dx[/tex]

and then it keeps looping
 
Last edited:
Physics news on Phys.org
If by "looping", you mean that you let [itex]dv= \frac{1}{x}[/itex] and [itex]u= e^x[/itex] and everything canceled out leaving you with
[tex]\int e^x lnx dx= \int e^x ln x dx[/itex]<br /> Of course!<br /> <br /> If you do integration by parts twice and the second time "swap" u and dv, that will always happen.<br /> <br /> Perhaps what you mean is that you again let [itex]dv= e^x[/itex] and [itex]u= \frac{1}{x}[/itex] so that you would up with<br /> [tex]e^x lnx dx= e^x lnx- \frac{1}{x}+ \int\frac{e^x}{x^2}dx[/itex]<br /> and continuing just gives you higher and higher powers in the denominator: Okay, that's good. What eventually happens is that, in the limit, you get ln x plus an infinite series. The integral cannot be done in terms of "elementary" functions.[/tex][/tex]
 
If, when solving by parts, you ever end up with your original integral, you can substitute a algebraically (like substitute the letter I for the origingal integral anywhere you see it) then just solve for I and that's your solution.

My professor broke solving by parts into three different types, and he called that one "around the world"
 
HallsofIvy said:
If by "looping", you mean that you let [itex]dv= \frac{1}{x}[/itex] and [itex]u= e^x[/itex] and everything canceled out leaving you with
[tex]\int e^x lnx dx= \int e^x ln x dx[/itex]<br /> Of course!<br /> <br /> If you do integration by parts twice and the second time "swap" u and dv, that will always happen.<br /> <br /> Perhaps what you mean is that you again let [itex]dv= e^x[/itex] and [itex]u= \frac{1}{x}[/itex] so that you would up with<br /> [tex]e^x lnx dx= e^x lnx- \frac{1}{x}+ \int\frac{e^x}{x^2}dx[/itex]<br /> and continuing just gives you higher and higher powers in the denominator: Okay, that's good. What eventually happens is that, in the limit, you get ln x plus an infinite series. The integral cannot be done in terms of "elementary" functions.[/tex][/tex]
[tex][tex] <br /> Except for the case [tex]\int_0^{\infty}dxe^{-ax}lnx=-\frac{\gamma+lna}{a}[/tex] which is exact.[/tex][/tex]
 

[tex]\text{ExpIntegralEi}[x] = - \int_{-x}^\infty \frac{e^{-t}}{t} dt[/tex]

Mathematica solution:
[tex]\int e^x ln{x} dx = \int_{-x}^\infty \frac{e^{-t}}{t} dt + e^x ln x[/tex]

---

[tex]\gamma[/tex] - EulerGamma (Euler-Mascheroni Constant)

[tex]\text{ExpIntegralEi}[z] = \sum_{k=1}^\infty \frac{z^k}{kk!} + \frac{1}{2} \left[ ln z - ln \left( \frac{1}{z} \right) \right] + \gamma[/tex]

[tex]\int_{-z}^\infty \frac{e^{-t}}{t} dt = \sum_{k=1}^\infty \frac{z^k}{kk!} + \frac{1}{2} \left[ ln z - ln \left( \frac{1}{z} \right) \right] + \gamma[/tex]

Orion1 solution:
[tex]\int e^x ln{x} dx = \sum_{k=1}^\infty \frac{x^k}{kk!} + \frac{1}{2} \left[ ln x - ln \left( \frac{1}{x} \right) \right] + e^x ln x + \gamma[/tex]

Reference:
http://functions.wolfram.com/GammaBetaErf/ExpIntegralEi/02/
[/Color]
http://mathworld.wolfram.com/Euler-MascheroniConstant.html
 
Last edited:
if "lnx" is the anle of any trignometric function. like intergration of "sinln dx "
 
ashu_manoo12 said:
if "lnx" is the anle of any trignometric function. like intergration of "sinln dx "

What? That made absolutely no sense whatsoever? ln(x) is the natural logarithm of x.