# Diff. eqn + erf (error function) (1 Viewer)

### Users Who Are Viewing This Thread (Users: 0, Guests: 1)

#### veronik

[SOLVED] Diff. eqn + erf (error function)

I’m stacked with this problem for many days, someone can help me pleeeeease:

(a) $$f \left( x \right) =\int _{-\infty }^{{x}^{2}/2}\!{e^{x-1/2\,{t}^{2 }}}{dt}$$

I foud the solution: $$f \left( x \right) =1/2\,{e^{x}}\sqrt {2\pi } \left( 1+{\it erf} \left( 1/4\,{x}^{2}\sqrt {2} \right) \right)$$

(b) Find the solution of the dfferential equatio:
$${\frac {d^{2}}{d{x}^{2}}}y \left( x \right) =f \left( x \right)$$ with y(0)=0 and dy(0)/dx = 0

In the form : $$y \left( x \right) =\int _{0}^{x}\! \left( x-t \right) f \left( t \right) {dt}$$

Veronica

#### tiny-tim

Homework Helper
Welcome to PF!

(b) Find the solution of the dfferential equatio:
$${\frac {d^{2}}{d{x}^{2}}}y \left( x \right) =f \left( x \right)$$ with y(0)=0 and dy(0)/dx = 0

In the form : $$y \left( x \right) =\int _{0}^{x}\! \left( x-t \right) f \left( t \right) {dt}$$

Veronica
Hi veronica! Welcome to PF!

(… the f in (a) isn't supposed to be the same as the f in (b), is it? …)

I suppose you know what $$\frac{d}{dx}\int _{0}^{x}\! \left( x-t \right) f \left( t \right) {dt}$$ is?

Are you sure they don't mean something like $$\int _{0}^{x}\! \left( x-t \right)^2 f \left( t \right) {dt}$$ ?

#### veronik

Hi veronica! Welcome to PF!

(… the f in (a) isn't supposed to be the same as the f in (b), is it? …)

I suppose you know what $$\frac{d}{dx}\int _{0}^{x}\! \left( x-t \right) f \left( t \right) {dt}$$ is?

Are you sure they don't mean something like $$\int _{0}^{x}\! \left( x-t \right)^2 f \left( t \right) {dt}$$ ?
1-the f in (a) is the same as in (b)...
2-I didnt calculate it..
3-it's x-t

its' (x-t).

#### tiny-tim

Homework Helper
3-it's x-t
No, it's not.

Try a simpler one: what is $$\frac{d}{dx}\int _{0}^{x}\! f\left( t \right) {dt}$$ ?

Hint: put f(t) = g´(t).

#### D H

Staff Emeritus
For (a), use the definition of the error function.

For (b), the solution is given. I'm a bit confused. What are you supposed to do?

EDIT:
You are given f(x). The information $d^2y/dx^2 = f(x)$ and $y(0) = y'(0) = 0$ are superfluous. Use the integral form.

Last edited:

#### veronik

I suppse ... the question is to demonstrate that the solution is in that form ( $$\int _{0}^{x}\! \left( x-t \right)^2 f \left( t \right) {dt}$$ )

tiny tin: the x-t was for you question about if it was in the form (x-t)f(t) or (x-t)^2f(t)

did u found the solution?

I'm getting tired :shy:... i'm from europe, it's already 2am here!

Last edited:

#### D H

Staff Emeritus
So, not the same f(x), as I thought originally. Use the Leibniz integral rule.

#### veronik

I think that it's the same f (it was stated like that in the question!)

#### D H

Staff Emeritus
The given equations are valid for any reasonable f(x). You need to show this is the case.

#### tiny-tim

Homework Helper
did u found the solution?

I'm getting tired :shy:... i'm from europe, it's already 2am here!
Hi Veronica!

I'm from London - we're one hour behind you! :zzz:

Look, with f(t) = g´(t), $$\frac{d}{dx}\int _{0}^{x}\! f\left( t \right) {dt}$$
$$=\,\frac{d}{dx}\int _{0}^{x}\! g'\left( t \right) {dt}$$
$$=\,\frac{d}{dx}\left[\,g\left(t)\,\right] _{0}^{x}$$
$$=\,\frac{d}{dx}(g(x)\,-\,g(0))$$
= g´(x) = f(x).

Now you try - what is $$\frac{d}{dx}\int _{0}^{x}\! x f\left( t \right) {dt}$$ ?

And then what is $${\frac {d^{2}}{d{x}^{2}}}\int _{0}^{x}\! \left( x-t \right) f \left( t \right) {dt}$$ ?

#### coomast

I’m stacked with this problem for many days, someone can help me pleeeeease:

(a) $$f \left( x \right) =\int _{-\infty }^{{x}^{2}/2}\!{e^{x-1/2\,{t}^{2 }}}{dt}$$

I foud the solution: $$f \left( x \right) =1/2\,{e^{x}}\sqrt {2\pi } \left( 1+{\it erf} \left( 1/4\,{x}^{2}\sqrt {2} \right) \right)$$

(b) Find the solution of the dfferential equatio:
$${\frac {d^{2}}{d{x}^{2}}}y \left( x \right) =f \left( x \right)$$ with y(0)=0 and dy(0)/dx = 0

In the form : $$y \left( x \right) =\int _{0}^{x}\! \left( x-t \right) f \left( t \right) {dt}$$

Veronica
The first part is indeed correct.

Hi Veronica!

I'm from London - we're one hour behind you! :zzz:

Look, with f(t) = g´(t), $$\frac{d}{dx}\int _{0}^{x}\! f\left( t \right) {dt}$$
$$=\,\frac{d}{dx}\int _{0}^{x}\! g'\left( t \right) {dt}$$
$$=\,\frac{d}{dx}\left[\,g\left(t)\,\right] _{0}^{x}$$
$$=\,\frac{d}{dx}(g(x)\,-\,g(0))$$
= g´(x) = f(x).

Now you try - what is $$\frac{d}{dx}\int _{0}^{x}\! x f\left( t \right) {dt}$$ ?

And then what is $${\frac {d^{2}}{d{x}^{2}}}\int _{0}^{x}\! \left( x-t \right) f \left( t \right) {dt}$$ ?
This is the way to show that the formula given is indeed the solution to the differential equation. However it does not give the derivation of the formula which (I assume) is to be found. The equation can be easily integrated to become:

$$y=\int \int f(x) dxdx+Ax+B$$

Naming now f(x)=g'(x) and g(x)=h'(x) and applying the boundary conditions you arrive at:

$$y(x)=h(x)-h(0)-xg(0)$$

This can be rewritten as:

[Begin Edit]
The solution has been removed.
I apologize if this caused any inconvenience, but I was under the assumption that it was not done to post the entire solution in the homework section, but is allowed for the other sections. Sorry about this. What if I was asked in a private message to give the solution?
[End Edit]

Which is the solution you are looking for. So, the boundary conditions are indeed necessary to arrive at the result.

Last edited:

#### D H

Staff Emeritus
Please do not give out complete solutions. In fact, please edit the solution out of that last post.

### The Physics Forums Way

We Value Quality
• Topics based on mainstream science
• Proper English grammar and spelling
We Value Civility
• Positive and compassionate attitudes
• Patience while debating
We Value Productivity
• Disciplined to remain on-topic
• Recognition of own weaknesses
• Solo and co-op problem solving