Applying Boundary Conditions at Infinity?



On Sun, 30 Jan 2011, Chris Nassar wrote:

> I'm trying to solve the differential equation A*p''(x)-p(x)/B=0 with the
> boundary conditions that p(infinity)=0 and p(xn)=pn.
> Maxima easily finds the general solution.
> ode2(A*'diff(p,x,2)-p/B,p,x);
> Is  A B  positive or negative?
> p
>                                    x                         x
>                               ---------------         - ---------------
>                               sqrt(A) sqrt(B)    sqrt(A) sqrt(B)
> (%o1)         p = %k1 %e                + %k2 %e
> Applying the boundary conditions does not give the expected result
> bc2(%o1,x=inf,p=0,x=xn,p=pn);
> Maxima treats infinity as just another variable.  Can I make maxima figure
> out that for p(infinity)=0 %k1 must be 0?  Since I know the answer I can
> just set %k1 to zero, but it does not seem like a satisfactory solution.
> I couldn't find anything in the manual about it.
> Any help would be appreciated.

No other takers?  Have you considered a change of independent variable
to something like u = 1-exp(xn-x) , to map the problem onto a finite
domain?

Now a question of my own: could one do something equivalent to this if
the homogeneous far-field condition were on the first derivative of p,
rather than on p itself?

-- 

HTH, and thanks,

Dan