Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

I found an integral calculated from what I understand with “differentation under the integration sign” method.

$$ \int_{-\infty}^{\infty}\exp\left(-b^{2}\left(x-c\right)^{2}\right)\mathrm{erf}\left(a\left(x-d\right)\right)\,\mathrm{d}x= {\frac{\sqrt\pi}{b}}\mathrm{erf}\left(\frac{ab\left(c-d\right)}{\sqrt{a^{2}+b^{2}}}\right) $$ for $b>0$.

The author of the original post explains how he obtained the solution:

“got it by differentiating the integrand w.r.t. a, then integrated over x=-inf..inf, then substituted a=sqrt(b*z)/sqrt(1-z) and integrated over z and then - most important - checked the result numerically.”

So I tried to follow that procedure and I get: $$I\left(a\right)=\int_{-\infty}^{\infty}\exp\left(-b^{2}\left(x-c\right)^{2}\right)\mathrm{erf}\left(a\left(x-d\right)\right)\,\mathrm{d}x$$

$$\frac{\mathrm{d}I\left(a\right)}{\mathrm{d}a}=\frac{\mathrm{d}}{\mathrm{d}a}\int_{-\infty}^{\infty}\exp\left(-b^{2}\left(x-c\right)^{2}\right)\mathrm{erf}\left(a\left(x-d\right)\right)\,\mathrm{d}x$$

$$\frac{\mathrm{d}I\left(a\right)}{\mathrm{d}a}=\int_{-\infty}^{\infty}\frac{\partial}{\partial a}\exp\left(-b^{2}\left(x-c\right)^{2}\right)\mathrm{erf}\left(a\left(x-d\right)\right)\,\mathrm{d}x$$

$$\frac{\mathrm{d}I\left(a\right)}{\mathrm{d}a}=\int_{-\infty}^{\infty}\frac{2\exp\left(-b^{2}(-c+x)^{2}-a^{2}(-d+x)^{2}\right)(-d+x)}{\sqrt{\pi}}\,\mathrm{d}x$$

$$\frac{\mathrm{d}I\left(a\right)}{\mathrm{d}a}=\frac{2b^{2}(c-d)}{\left(a^{2}+b^{2}\right)^{3/2}}\exp\left(-\frac{a^{2}b^{2}(c-d)^{2}}{a^{2}+b^{2}}\right)$$

Now I substitute $z=\frac{a^2}{a^2+b^2}$ and after some manipulations I get the right side of the last equation: $$ 2\sqrt{b}(c-d)(1-z)^{3/2}\exp\left(-zb^{2}(c-d)^{2}\right) $$

I’m not sure what my next step should be so I would appreciate any suggestions.

share|improve this question

1 Answer 1

up vote 2 down vote accepted

Let's try a close variant : differentiate relatively to $d$ instead of $a$ at the start (to avoid the additional factor $(x-d)$ in the integral and complications in the final integration) :

$$I\left(d\right)=\int_{-\infty}^{\infty}e^{-b^{2}\left(x-c\right)^{2}}\ \mathrm{erf}\left(a\left(x-d\right)\right)\,\mathrm{d}x$$

$$\frac{\mathrm{d}I\left(d\right)}{\mathrm{d}d}=\frac {-2a}{\sqrt{\pi}}\int_{-\infty}^{\infty}e^{-b^{2}(x-c)^{2}-a^{2}(x-d)^{2}}\,\mathrm{d}x$$

$$\frac{\mathrm{d}I\left(d\right)}{\mathrm{d}d}=\frac {-2a}{\sqrt{\pi}}\int_{-\infty}^{\infty}e^{-b^{2}(x-c)^{2}-a^{2}(x-d)^{2}}\,\mathrm{d}x$$

$$\frac{\mathrm{d}I\left(d\right)}{\mathrm{d}d}=\frac {-2a}{\sqrt{\pi}}\int_{-\infty}^{\infty}e^{-\left(a^2+b^2\right)\left(x-\frac{b^2c+a^2d}{a^2+b^2}\right)^2+\frac{\left(b^2c+a^2d\right)^2}{a^2+b^2}-\left(b^2c^2+a^2d^2\right)}\,\mathrm{d}x$$

$$\frac{\mathrm{d}I\left(d\right)}{\mathrm{d}d}=\frac {-2a}{\sqrt{\pi}}e^{\frac{\left(b^2c+a^2d\right)^2}{a^2+b^2}-\left(b^2c^2+a^2d^2\right)}\int_{-\infty}^{\infty}e^{-\left(a^2+b^2\right)y^2}\,\mathrm{d}y$$

$$\frac{\mathrm{d}I\left(d\right)}{\mathrm{d}d}=\frac {-2a}{\sqrt{\pi}}e^{\frac{-a^2b^2\left(c-d\right)^2}{a^2+b^2}}\frac{\sqrt{\pi}}{\sqrt{a^2+b^2}}=-2a\frac{e^{\frac{-a^2b^2\left(c-d\right)^2}{a^2+b^2}}}{\sqrt{a^2+b^2}}$$

At this point we have to integrate again relatively to $d$ to get (up to a function $C$ independent of $d$) :

$$I\left(d\right)=\frac {\sqrt{\pi}}b\operatorname{erf}\left(\frac{ab(c-d)}{\sqrt{a^2+b^2}}\right)+C(a,b,c)$$

(Alpha integration check : note that the denominator is $b$ and not $\sqrt{b}$ nor my earlier $ab$ as I checked numerically!)

After that you'll just have to prove that $C(a,b,c)\equiv 0$

Note that the integration relatively to $d$ seems more straightforward than relatively to $a$ in your case (I'm not saying it can't be done your way!).

Hoping it clarified things a little even if it didn't answer your question,

share|improve this answer
    
Thanks a lot! It explains all I wanted to know. And yes, I made a typo in the first post: it should be $\frac{\sqrt{\pi}}{b}$. –  petru Apr 29 '12 at 18:58
    
@petru: I am glad it helped! Fine continuation (I noticed your other integral), –  Raymond Manzoni Apr 29 '12 at 19:05

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.