How to integrate an ordinary differential equation around a singular point?

Click For Summary

Discussion Overview

The discussion revolves around the integration of a second-order ordinary differential equation (ODE) with a singular point at the origin. The equation is given as f'' + (E - A(A+1)/x^2)f = 0, with boundary conditions specified. Participants explore numerical methods for solving the ODE, particularly near the singularity at x=0.

Discussion Character

  • Exploratory
  • Technical explanation
  • Mathematical reasoning

Main Points Raised

  • One participant presents the ODE and boundary condition, noting that f is on the order of x^(A+1) near the origin.
  • Another participant suggests using the 4th-order Runge-Kutta method for numerical integration but expresses concern about the dominance of the A(A+1)/x^2 term near the origin and the potential error control issues.
  • A third participant proposes a specific analytical solution involving Bessel functions, but later clarifies that they are more interested in a general numerical method for handling singular points rather than the analytical solution provided.

Areas of Agreement / Disagreement

Participants do not reach a consensus on the best numerical method to apply near the singular point, and there are differing views on the applicability of the Runge-Kutta method in this context.

Contextual Notes

The discussion highlights concerns about the behavior of the ODE near the singularity and the challenges of numerical integration in such regions, but does not resolve these issues or provide a definitive method.

Who May Find This Useful

Researchers or students interested in numerical methods for solving ordinary differential equations, particularly those with singular points, may find this discussion relevant.

wdlang
Messages
306
Reaction score
0
i have a second order ordinary differential equation of f(x):

f''+(E-A(A+1)/x^2)f=0, where A is a positive integer, E is a real constant

the domain is [0, \infty).

the boundary condition is f(x=0)=0

since this is a linear equation, i only need to determine f up to a overall constant

how to do this numerically from the origin and outward?

we can prove that in the neighborhood of the origin, f is on the order of x^(A+1)
 
Physics news on Phys.org
a naive idea is to use the boundary condition mentioned above and the 4th-order Runge-Kutta method to do it

my concern is that in near the origin, the A(A+1)/x^2 term is too large

and therefore i have no idea whether the RK method will work, or if it does, how to control the error by shrinking the step length.
 
If I do not mistaken, the solution

[tex]f(x) = C\sqrt{x}J_{A+1/2}(\sqrt{E}x)[/tex]

where C is an arbitrary constant fit your ODE and the boundary condition f(x=0)=0.
 
kosovtsov said:
If I do not mistaken, the solution

[tex]f(x) = C\sqrt{x}J_{A+1/2}(\sqrt{E}x)[/tex]

where C is an arbitrary constant fit your ODE and the boundary condition f(x=0)=0.

thanks a lot

however, i am not interested in the analytical solution

i want a general numerical method to deal with such singular points

the equation above is just an example, for which, luckily, an analytical solution is available
 

Similar threads

  • · Replies 5 ·
Replies
5
Views
4K
  • · Replies 1 ·
Replies
1
Views
3K
  • · Replies 2 ·
Replies
2
Views
3K
  • · Replies 1 ·
Replies
1
Views
4K
  • · Replies 1 ·
Replies
1
Views
3K
  • · Replies 0 ·
Replies
0
Views
4K
  • · Replies 7 ·
Replies
7
Views
3K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 3 ·
Replies
3
Views
4K
  • · Replies 2 ·
Replies
2
Views
4K