Finding a solution to this equation using Frobenius method

  • Context: Graduate 
  • Thread starter Thread starter climbon
  • Start date Start date
  • Tags Tags
    Frobenius Method
Click For Summary
SUMMARY

The discussion focuses on solving the differential equation y'' + (1/x)y' + [(x^2) + k + (m^2 / x^2)]y = 0 using the Frobenius method. The user derives a recurrence relation a_(n+4) = [-ka_(n+2) - a_(n)] / [n^2 +/- 2inm] but encounters difficulties in progressing further. They provide Mathematica code to validate their solution, which involves calculating coefficients a_n and b_n, and ultimately expresses the solution as y(x)=K_1 ∑ a_n x^(n+mi) + K_2 ∑ b_n x^(n-mi). The user believes that the imaginary components will cancel out under real initial conditions.

PREREQUISITES
  • Understanding of differential equations, specifically second-order linear equations.
  • Familiarity with the Frobenius method for solving differential equations.
  • Proficiency in Mathematica for numerical solutions and plotting functions.
  • Knowledge of complex numbers and their role in differential equations.
NEXT STEPS
  • Study the Frobenius method in greater detail, focusing on convergence criteria.
  • Learn about the application of complex analysis in solving differential equations.
  • Explore advanced features of Mathematica for solving differential equations numerically.
  • Investigate the implications of initial conditions on the solutions of differential equations.
USEFUL FOR

Mathematicians, physicists, and engineers working with differential equations, particularly those interested in series solutions and numerical methods using Mathematica.

climbon
Messages
16
Reaction score
0
Hi, I have this equation to solve.

y'' + (1/x)y' + [(x^2) + k + (m^2 / x^2)]y = 0

now, I've tried to solve this using frobenius method but cannot formulate a solution.

I have that

a_(n+4) = [-ka_(n+2) - a_(n)] / [n^2 +/- 2inm]

is my recurrence relation, but now I'm stuck and don't know how to proceed, any help greatly appreciated.

thanks.
 
Physics news on Phys.org
Using the standard convention of letting a_0=1, I get:

a_1=0
a_2=-\frac{h a_0}{(2+c)(1+c)+1+m^2}

with the remaining odd a_n=0 and even a_n equal to:

a_n=-\frac{a_{n-2}+a_{n-4}}{(n+c)(n+c+1)+1+m^2}

with a similar relation for the other root expressed in b_n and so I can write the solution as:

y(x)=K_1 \sum_{n=0}^{\infty} a_n x^{n+mi}+K_2 \sum_{n=0}^{\infty} b_n x^{n-mi},\quad x>0

and I believe because of the conjugates in the solution, for real initial conditions y(x_0)=y_0, and y'(x_0)=y_1, the imaginary component is annihilated leaving the desired real solution.

If you into it, here's some Mathematica code to check the solution. I believe it's correct but not as close as I would have expected. May be some convergence issues though.

Code:
nmax = 75; 
x0 = 0.1; 
y0 = 0; 
y1 = 1.; 
h = 5; 
m = 4; 

mysol = NDSolve[{x^2*Derivative[2][y][x] + x*Derivative[1][y][x] + (x^4 + h*x^2 + m^2)*y[x] == 0, y[x0] == y0, 
     Derivative[1][y][x0] == y1}, y, {x, x0, 2}]; 

sol[x_] := Evaluate[y[x] /. mysol]; 
p1 = Plot[y[x] /. mysol, {x, x0, 2}]; 

c1 = I*m; 
Subscript[a, 0] = 1; 
Subscript[a, 1] = 0; 
Subscript[a, 2] = ((-h)*Subscript[a, 0])/((2 + c1)*(1 + c1) + 1 + m^2); 
Subscript[a, 3] = 0; 

Table[Subscript[a, n] = -(Subscript[a, n - 4] + h*Subscript[a, n - 2])/((n + c1)*(n + c1 - 1) + 1 + m^2), 
   {n, 4, nmax}]; 

f1[x_] := Sum[Subscript[a, n]*x^(n + c1), {n, 0, nmax}]
f1d[x_] = D[f1[x], x]; 

c2 = (-I)*m; 
Subscript[b, 0] = 1; 
Subscript[b, 1] = 0; 
Subscript[b, 2] = ((-h)*Subscript[b, 0])/((2 + c2)*(1 + c2) + 1 + m^2); 
Subscript[b, 3] = 0; 

Table[Subscript[b, n] = -((Subscript[b, n - 4] + h*Subscript[b, n - 2])/((n + c2)*(n + c2 - 1) + 1 + m^2)), 
   {n, 4, nmax}]; 

f2[x_] := Sum[Subscript[b, n]*x^(n + c2), {n, 0, nmax}]; 
f2d[x_] = D[f2[x], x]; 

myks = First[NSolve[{y0 == k1*f1[x0] + k2*f2[x0], y1 == k1*f1d[x0] + k2*f2d[x0]}, {k1, k2}]]; 

myy[x_] := k1*f1[x] + k2*f2[x] /. myks; 

myd1[x_] = D[myy[x], x]; 
myd2[x_] = D[myy[x], {x, 2}]; 

N[x^2*myd2[x] + x*myd1[x] + (x^4 + x^2*h + m^2)*myy[x]] /. x -> 0.334

p2 = Plot[myy[x], {x, x0, 2}, PlotStyle -> Red]
Show[{p1, p2}]
 
Last edited:

Similar threads

  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 1 ·
Replies
1
Views
930
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 2 ·
Replies
2
Views
6K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 2 ·
Replies
2
Views
3K
  • · Replies 4 ·
Replies
4
Views
3K
  • · Replies 5 ·
Replies
5
Views
16K
  • · Replies 3 ·
Replies
3
Views
2K