Green's Function for Two-Point Boundary Value Problem

  • Thread starter Thread starter Brian T
  • Start date Start date
Brian T
Messages
130
Reaction score
31

Homework Statement


Determine the Green's function for the following problem
$$-u'' + cu = f$$ in the domain D = (0,1) with BC's u(0)=u(1)=0

Homework Equations


The Green's function is given by
$$G(x,y) =\begin{cases}
\frac{1}{W} u_0(x)u_1(y), &\text{for } 0 \leq y \leq x \leq 1, \\
\frac{1}{W} u_0(x)u_1(y), &\text{for } 0 \leq x \leq y \leq 1.
\end{cases} $$
Where W is the Wronskian of the solutions to the homogeneous equation u_0, u_1, where u_0, u_1 satisfy the boundary conditions u_0(0)=1, u_0(1)=0, u_1(0)=0, u_1(1)=1

The Attempt at a Solution


I am able to find the Green's function assuming c = const.
To find the fundamental solutions, we solve the homogenous equation -u'' + cu = 0.
The fundamental solutions are e^{\sqrt{c}x}, e^{-\sqrt{c}x}. To satisfy the boundary conditions, choose the appropriate linear combination of these fundamental solutions. After a little algebra, I get:
$$u_0(x) = \frac{1}{1 - e^{2\sqrt{c}}}e^{\sqrt{c}x} + \frac{-e^{2\sqrt{c}}}{1-e^{2\sqrt{c}}))}e^{-\sqrt{c}x}$$
$$u_1(x) = \frac{1}{e^{\sqrt{c}} - e^{-\sqrt{c}}}e^{\sqrt{c}x} + \frac{1}{e^{-\sqrt{c}} - e^{\sqrt{c}}}e^{-\sqrt{c}x}$$

I know this is all correct (plugged in and verified my solutions).
The question I have is for c not a constant, i.e. c = c(x). Since the problem does not specify whether c is constant or a function, I would assume that it could be a function... The DE becomes
$$ -u'' + c(x)u = f $$
This would require getting the fundamental solutions by solving
$$ -u'' + c(x)u = 0 $$
which I am unsure of how to do for an arbitrary function c(x). There are certain methods for certain types of functions of c(x), but does anyone have a clue on how to solve this ODE for a general c(x)? If not do you think the problem could be referring to a constant c (in the actual chapter the author usually takes the coefficient to be an arbitrary function)? By the way, the book is numerical PDE's by Larrson
 
Physics news on Phys.org
There is no general way of solving that for any function c(x). By the way, finding Green's function for the case c constant can by simplified by using cosh(cx) and sinh(cx) rather than e^{cx} and e^{-cx} .
 
  • Like
Likes Brian T
HallsofIvy said:
There is no general way of solving that for any function c(x). By the way, finding Green's function for the case c constant can by simplified by using cosh(cx) and sinh(cx) rather than e^{cx} and e^{-cx} .

Ohh alright, thanks. I'll try that instead :).
 
There are two things I don't understand about this problem. First, when finding the nth root of a number, there should in theory be n solutions. However, the formula produces n+1 roots. Here is how. The first root is simply ##\left(r\right)^{\left(\frac{1}{n}\right)}##. Then you multiply this first root by n additional expressions given by the formula, as you go through k=0,1,...n-1. So you end up with n+1 roots, which cannot be correct. Let me illustrate what I mean. For this...
Back
Top