Solving Difference Equations with Homogeneous and Inhomogeneous Parts?

Click For Summary

Discussion Overview

The discussion revolves around solving a specific difference equation that includes both homogeneous and inhomogeneous components. Participants explore various methods and approaches to find solutions, addressing the roles of parameters and initial conditions in the context of difference equations.

Discussion Character

  • Technical explanation
  • Mathematical reasoning
  • Debate/contested

Main Points Raised

  • One participant presents a difference equation and seeks guidance on solving it, noting its complexity due to the presence of two unknowns, n and p.
  • Another participant suggests finding the general solution to the homogeneous part and a particular solution for the inhomogeneous part, indicating that the overall solution is a combination of both.
  • A different viewpoint emphasizes treating p as a parameter rather than an unknown, proposing a method to find roots of unity related to the equation.
  • Subsequent posts discuss the substitution of specific forms into the homogeneous equation and the implications of the roots obtained, with some participants expressing uncertainty about the next steps.
  • Participants clarify the process of obtaining solutions for the homogeneous equation and the necessity of adding a particular solution to address the inhomogeneous part.
  • One participant attempts to derive a specific solution by substituting values for p and explores the resulting roots, while another suggests a constant as a particular solution.
  • There is a discussion about the role of coefficients and the nature of solutions, with some participants questioning the correctness of their approaches and seeking clarification on the terms involved.

Areas of Agreement / Disagreement

Participants express differing views on the treatment of parameters and the methods for solving the equation. While some agree on the general approach of separating homogeneous and inhomogeneous solutions, others raise questions about specific steps and interpretations, indicating that the discussion remains unresolved in certain aspects.

Contextual Notes

Participants note various assumptions and conditions, such as the dependence on the parameter p and the need for initial conditions to determine specific coefficients. The discussion includes unresolved mathematical steps and differing interpretations of the solution process.

FeDeX_LaTeX
Science Advisor
Messages
436
Reaction score
13
Hello;

This is not a homework question, but something I was wondering about solving difference equations.

For example, how would I solve the following difference equation;

[tex]F_{n} = 2F_{n - 2p + 5} + 6p - 17; n, p \in \mathbb N[/tex]

Since it has homogeneous and inhomogeneous parts together (and two unknowns - n and p) I don't know what to do. Any help on how I would solve this would be appreciated, thanks.

(For where I got this formula from - I found this when thinking about a formula that would give me the minimum number of moves to complete a game of Tower of Hanoi with n discs and p poles.)
 
Last edited:
Physics news on Phys.org
Since that is a linear equation, you can do this:
Find the general solution to the homogeneous equation [itex]F_n= F_{n- 2p+ 5}[/itex]. call that "[itex]F_h[/itex]".

Find a single function, [itex]F_p[/itex], that solves the entire equation.

The general solution to the entire equation is the sum, [itex]F_h+ F_p[/itex].
 
For every p, your equation is just an ordinary difference equation. So p is not really an unknown, rather a parameter.
Define [tex]k\equiv |2p-5|[/tex], find the k roots of unity

[tex]\lambda^k=1\qquad\rightarrow \lambda_m=e^{i2m\pi/k},\qquad m=0,1,\dots,k-1[/tex]

Then the solution is

[tex]F_n=\sum_{m=1}^{k-1}c_m\lambda_{m}^{n}+c+dn[/tex]

The coeeficients c and d are really just one coefficient, because you get a linear relationship between them when you substitute the solution in the equation. To get the other coefficients you need initial conditions.
 
Last edited:
Sorry, I left out the co-efficient 2. It should read;

[tex]F_{n} = 2F_{n - 2p + 5} + 6p - 17; n, p \in \mathbb N[/tex]

I'm not sure I completely understand what to do; is this working correct?

Let [tex]F_{n} = k^{n}[/tex]

[tex]\Rightarrow k^{n} = 2k^{n - 2p + 5}[/tex]

[tex]\Rightarrow k^{2p - 5} = 2[/tex]

[tex]\Rightarrow k = \sqrt[2p - 5] {2}e^{\frac{2im\pi}{2p - 5}}[/tex] for [tex]p \geq 3[/tex].

Where do I go from here? I think I have to put the first root in like so;

[tex]F_{n} = A\left(\sqrt{2}e^{\frac{2i\pi}{2p - 5}}\right)^{n} + B(\cdots)^{n}[/tex]

But what do I put in the "B" bracket? Where does the 'm' come from, and what does it do/why is it there?
 
FeDeX_LaTeX said:
But what do I put in the "B" bracket? Where does the 'm' come from, and what does it do/why is it there?

You have to substitute

[tex]F_n=\lambda^n[/tex]

in the omogeneous equation, not in the complete one.
You get a polynomial in lambda, that will give various roots, and each of them is a solution. Since the equation is linear, you can multiply each solution by an arbitrary constant and then sum up everything, that's why the "m".
This way you get a solution of the homogeneous equation. To get a complete solution, you have to add another term, that in your particular example must be simply a constant.
 
Petr Mugver said:
You have to substitute

[tex]F_n=\lambda^n[/tex]

in the omogeneous equation, not in the complete one.

But isn't that what I've done by saying that [tex]F_{n} = k^{n}[/tex] for the homogeneous part?

You get a polynomial in lambda, that will give various roots, and each of them is a solution.

Do you mean this? [tex]\sqrt[2p - 5] {2}e^{\frac{2im\pi}{2p - 5}}[/tex] I think it gives the roots for k, but I'm not entirely sure what to do from there except to just assume that whole thing is one root for k.

Since the equation is linear, you can multiply each solution by an arbitrary constant and then sum up everything, that's why the "m".

Okay, that makes sense. I understand this.

This way you get a solution of the homogeneous equation. To get a complete solution, you have to add another term, that in your particular example must be simply a constant.

Is the term that I have to add (17 - 6p)?

Because if we say that the homogeneous part is t, then t = 2t + 6p - 17, so t = 17 - 6p? Or not?
 
Okay let' suppose p=4. Then your equation becomes

[tex]F_n-2F_{n-3}=1[/tex]

Let's put

[tex]F_n=\lambda^n[/tex]

Substituting in the homogeneous equation tou get

[tex]\lambda^3=2[/tex]

and you have 3 roots:

[tex]\lambda_1=2^{1/3}\qquad \lambda_2=2^{1/3}e^{2\pi i/3}\qquad \lambda_3=2^{1/3}e^{-2\pi i/3}[/tex]

So the solution of the homogeneous eq is

[tex]F^{Omog}_n=c_1\lambda_1^n+c_2\lambda_2^n+c_3\lambda_3^n[/tex]

Now we need a particular solution of the inhomogeneous equation. Let's try

[tex]F^{Part}_n=constant[/tex]

Substituting, we find constant=-1, so finally

[tex]F_n=c_1\lambda_1^n+c_2\lambda_2^n+c_3\lambda_3^n-1[/tex]
 
Last edited:
Thanks. I think I know how to solve it now.
 

Similar threads

  • · Replies 36 ·
2
Replies
36
Views
6K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 3 ·
Replies
3
Views
4K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 1 ·
Replies
1
Views
4K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 9 ·
Replies
9
Views
1K
  • · Replies 3 ·
Replies
3
Views
3K