Originally Posted by

**kingwinner** Maybe my background is just weak...I was thinking about this for almost 1.5 hours already, but I still end up totally confused. Perhaps this is because I was never able to understand the ideas of a function and change of variables completely...perhaps I have a serious conceptual flaw.

========================================

**Consider the following partial differential equation with initial values(IV) and boundary conditions(BC):**

**u_t - k u_xx = x + 2t, 1<x<7, t>0**

**BC: u(1,t) = u(7,t) = 0**

**IV: u(x,0) = x+5**

**Our goal is to transform the above to the interval [0,6].**

**Let w = x-1.**

**Transform the whole problem to the interval w E [0,6]. (write in terms of w)**

================================================

u_x = u_w dw/dx = (u_w) (1) = u_w

u_xx = ...(apply chain rule again) = u_ww

[On the left side, think of u as u(x,t). On the right side, think of u as u(w,t)]

BC:

x=1 <=> w=0

x=7 <=> w=6

So the boundary conditions get transformed to u(0,t)=u(6,t)=0 [here think of u as u(w,t)]

IV:

We know u(x,0) = x+5

=> u(w,0) = w+5

[I believe the logic in this step cannot be wrong, consider e.g. f(4z)=cos(4z), now how do we find f(4z-y)? Of course, f(4z-y)=cos(4z-y). How do we find f(z)? Surely, f(z)=cos(z). Right??]

So my final answer is: [here think of u as u(w,t)]

u_t - k u_ww = w+1+2t, 0<w<6, t>0

BC: u(0,t) = u(6,t) = 0

IV: u(w,0) = w+5

However, I really have some bad feeling that the result u(w,0) = w+5 is wrong, but I don't know where the mistake is.

I tried to calculate it in a different way and the answer is the same.

u(w,0)

= u(x-1,0) = (x-1)+5

=> u(w,0) = w+5

Can someone please kindly explain why and where my mistake is? What is the correct answer?

Any help is greatly appreciated!

[also under discussion in sos math cyberboard]