- #1
bolbteppa
- 309
- 41
Following Butkov, a second order ode
[tex]A(x)y'' + B(x)y' + C(x)y = D(x)[/tex]
can always be brought into Sturm-Liouville form
[tex]\tfrac{d}{dx}[p(x)y'] - s(x)y = f(x)[/tex]
after multiplying across by
[tex]H(x) = - \tfrac{1}{A(x)}e^{\int^x \tfrac{B(t)}{A(t)}dt}.[/tex]
He then says the function [itex]s(x)[/itex] can "often" be written as
[tex]s(x) = s_0(x) - \lambda r_0(x)[/tex]
where [itex]0 \leq s_0(x)[/itex], [itex]0 \leq r_0(x)[/itex] & [itex]\lambda[/itex] is fixed.
I just don't see how once can be comfortable with this or how one can use a statement like this in general. How does one take a general second order ode & concretely turn it into something involving [itex]\lambda[/itex]?
For instance, in the case that [itex]A[/itex], [itex]B[/itex] & [itex]C[/itex] are polynomials of degree [itex]2[/itex], [itex]1[/itex] & [itex]0[/itex]:
[tex](ax^2+bx+c)y'' + (dx + e)y' + fy = F(x)[/tex]
I can see that [itex]f[/itex] will be an eigenvalue after multiplication by [itex]H(x)[/itex], but how would you deal with a case like
[tex](ax^2+bx+c)y'' + (dx + e)y' + \sin(x)y = F(x)[/tex]
[tex]A(x)y'' + B(x)y' + C(x)y = D(x)[/tex]
can always be brought into Sturm-Liouville form
[tex]\tfrac{d}{dx}[p(x)y'] - s(x)y = f(x)[/tex]
after multiplying across by
[tex]H(x) = - \tfrac{1}{A(x)}e^{\int^x \tfrac{B(t)}{A(t)}dt}.[/tex]
He then says the function [itex]s(x)[/itex] can "often" be written as
[tex]s(x) = s_0(x) - \lambda r_0(x)[/tex]
where [itex]0 \leq s_0(x)[/itex], [itex]0 \leq r_0(x)[/itex] & [itex]\lambda[/itex] is fixed.
I just don't see how once can be comfortable with this or how one can use a statement like this in general. How does one take a general second order ode & concretely turn it into something involving [itex]\lambda[/itex]?
For instance, in the case that [itex]A[/itex], [itex]B[/itex] & [itex]C[/itex] are polynomials of degree [itex]2[/itex], [itex]1[/itex] & [itex]0[/itex]:
[tex](ax^2+bx+c)y'' + (dx + e)y' + fy = F(x)[/tex]
I can see that [itex]f[/itex] will be an eigenvalue after multiplication by [itex]H(x)[/itex], but how would you deal with a case like
[tex](ax^2+bx+c)y'' + (dx + e)y' + \sin(x)y = F(x)[/tex]