Definite integral of exp and error function

petru

Hi,

I've been trying to evaluate the following integral

[tex] \int_{z}^{\infty}\exp\left(-y^{2}\right)\mathrm{erf}\left(b\left(y-c\right)\right)\,\mathrm{d}y [/tex]

or equivalently

[tex] \int_{z}^{\infty}\exp\left(-y^{2}\right)\mathrm{erfc}\left(b\left(y-c\right)\right)\,\mathrm{d}y [/tex]

[tex]\mathrm{erf}\left(x\right)=\frac{2}{\sqrt{\pi}} \int_{0}^{x}\exp\left(-u^{2}\right)\,\mathrm{d}u, \quad\quad \mathrm{erfc}\left(x\right)=1-\mathrm{erf}\left(x\right)=\frac{2}{\sqrt{\pi}} \int_{x}^{+\infty}\exp\left(-u^{2}\right)\,\mathrm{d}u[/tex]

I guess I tried to employ all techniques I'm familiar with but with no result.
Can anyone help me with this one, please?
Thank you!

mathman

To the best of my knowledge, it can't be done analytically. I suggest you start with the erf representation and see if the two exponentials might be combined into one, so that you might have an erf for the integral.

petru

Thanks mathman for your reply. I guess I'm not able to deal with this integral. I have a question though. I'm not a mathematician nor a math student so I was wondering if anyone could explain to me why the integral

[tex]\int_{-\infty}^{\infty}\exp\left(-y^{2}\right) \mathrm{erf}\left(b\left(y-c\right)\right)\,\mathrm{d}y=-\sqrt{\pi}\,\mathrm{erf}\left(\frac{bc}{\sqrt{1+b^{2}}}\right)[/tex]

can be evaluated quite easily (using differentiation under integral sign method) and the integral from my original post seems to have no analytical solution?

Thanks!

mathman

I haven't looked at it in detail, but it looks like the problem is analogous to integrating the Gaussian. When you integrate over the entire real line you get a neat analytic solution, but integrating over part of the line ends up with erf.

petru

Ok, I guess I know what you mean. Thanks again!

JJacquelin

Hi !

in attachment, a method for solving the definite integral.

petru

Hi JJacquelin! Your post helped me with showing that the constant of integration [tex]C=0[/tex] in a more general formula:

[tex]
\int_{-\infty}^{\infty}\exp\left(-b^{2}(x-c)^{2}\right)\mathrm{erf}\left(a\left(x-d\right)\right)\,\mathrm{d}x=\frac{\sqrt{\pi}}{b} \mathrm{erf}\left(\frac{ab\left(c-d\right)}{\sqrt{a{}^{2}+b{}^{2}}}\right),\quad b>0
[/tex]

Thank you!

Similar Threads for: Definite integral of exp and error function
Thread	Forum	Replies
Definite Integral of Trig Function	Calculus & Beyond Homework	2
definite integral of complex gaussial-like function	Calculus	17
Definite integral of rational function	Calculus & Beyond Homework	6
Definite integral of an even function	Calculus & Beyond Homework	3
Definite Integral of ln*algebraic function	Introductory Physics Homework	21

mathman Recognitions: Science Advisor	To the best of my knowledge, it can't be done analytically. I suggest you start with the erf representation and see if the two exponentials might be combined into one, so that you might have an erf for the integral.

mathman Recognitions: Science Advisor	Definite integral of exp and error function I haven't looked at it in detail, but it looks like the problem is analogous to integrating the Gaussian. When you integrate over the entire real line you get a neat analytic solution, but integrating over part of the line ends up with erf.

JJacquelin	Hi ! in attachment, a method for solving the definite integral. Attached Thumbnails

petru	Hi JJacquelin! Your post helped me with showing that the constant of integration [tex]C=0[/tex] in a more general formula: [tex] \int_{-\infty}^{\infty}\exp\left(-b^{2}(x-c)^{2}\right)\mathrm{erf}\left(a\left(x-d\right)\right)\,\mathrm{d}x=\frac{\sqrt{\pi}}{b} \mathrm{erf}\left(\frac{ab\left(c-d\right)}{\sqrt{a{}^{2}+b{}^{2}}}\right),\quad b>0 [/tex] Thank you!