Solve y'''=y'': Using Power Series to Solve Non-Linear Diff Eq

  • Context: Graduate 
  • Thread starter Thread starter lkj-17
  • Start date Start date
Click For Summary

Discussion Overview

The discussion focuses on solving the non-linear differential equation y''' = y'' using power series and integration techniques. Participants explore various methods to derive solutions, including assumptions about initial conditions and the use of integrals involving the Exponential Integral function.

Discussion Character

  • Exploratory
  • Mathematical reasoning
  • Technical explanation

Main Points Raised

  • One participant inquires about the use of power series to solve the equation.
  • Another participant proposes assuming initial conditions y(0) = A, y'(0) = B, and y''(0) = C, and derives a series expansion for y(x) based on these assumptions.
  • A different approach is suggested, where the equation is rearranged to y''' = y''/y' and integrated to yield y'' = κ + ln(y'), introducing constants of integration.
  • This participant further discusses the integration of the resulting expression and references the Exponential Integral function, providing a formula for y' and suggesting numerical techniques for obtaining values for y(x).
  • A correction is made regarding the expression for y(x), clarifying the integration constant used in the formula.

Areas of Agreement / Disagreement

Participants present multiple competing views on how to approach the solution, with no consensus reached on a single method or final expression for y(x).

Contextual Notes

Limitations include dependence on initial conditions and the complexity of integrating expressions involving the Exponential Integral function, which remains unresolved.

lkj-17
Messages
5
Reaction score
0
How to use power series to solve this non-linear differential equation?
 
Physics news on Phys.org
First, because this is a third order equation, its general solution will involve 3 undetermined constants. So, assume y(0)= A, y'(0)= B, y"(0)= C.

From y'y"'= y", y"'= y"/y' and so y"'(0)= C/B. Now, differentiating y"'= y"/y' again, yiv= (y'y"'- y"2)/y'2 so yiv(0)= (B(C/B)- C^2)/B2= (C- C2)/B2.

So far, we have y(x)= y(0)+ y'(0)x+ (1/2)y"(0)x2+ (1/3!)y"'(0)x3+ (1/4!)yiv(0)x4+ ...= A+ Bx+ (1/2)Cx2+ (1/6)(C/B)x3+ (1/24){(C- C2)/B2}x^4+ ...

Continue like that to get higher terms.
 
Re-arrange your equation as

[tex]y^{\prime \prime \prime} = \frac{y^{\prime \prime}}{y^{\prime}}[/tex]

Now integrate with respect to x to get

[tex]y^{\prime \prime} = \kappa + \ln{y^{\prime}}[/tex]

where [tex]\kappa[/tex] is a constant of integration. Now re-arrange and integrate to get

[tex]\int{\frac{d y^{\prime}}{\kappa + \ln{y^{\prime}}} = x + \epsilon[/tex]

where [tex]\epsilon[/tex] is another constant.

I checked the Integrator (Wolfram site), and it gave the integral as:

[tex]\int{\frac{dw}{a + \ln{w}}} = e^{-a}Ei(a + \ln{w})[/tex]

where [tex]Ei[/tex] is the Exponential Integral.

Hence we can apply this to our integral to get

[tex]e^{- \kappa} Ei(\kappa + \ln{y^{\prime}) = x + \epsilon[/tex]

which we can re-arrange as

[tex]y^{\prime} = exp(Ei^{-1}(e^{\kappa}(x + \epsilon)) - \kappa)[/tex]

We can tidy this up a bit by making

[tex]e^{\kappa} = \alpha[/tex]

and

[tex]\epsilon \alpha = \beta[/tex]

and then integrate to get

[tex]y(x) = \frac{1}{\beta}\int{exp(Ei^{-1}(\alpha x + \beta)) dx}[/tex]

and then you'll have to try some numerical techniques to obtain values for y(x) (I have to idea how to express the inverse of "Ei".)
 
Last edited:
Sorry that should read:

[tex]y(x) = \frac{1}{\alpha}\int{exp(Ei^{-1}(\alpha x + \beta)) dx}[/tex]
 

Similar threads

  • · Replies 5 ·
Replies
5
Views
3K
  • · Replies 12 ·
Replies
12
Views
2K
  • · Replies 3 ·
Replies
3
Views
3K
  • · Replies 2 ·
Replies
2
Views
3K
  • · Replies 9 ·
Replies
9
Views
3K
  • · Replies 8 ·
Replies
8
Views
4K
  • · Replies 7 ·
Replies
7
Views
3K
  • · Replies 7 ·
Replies
7
Views
4K
  • · Replies 1 ·
Replies
1
Views
959
  • · Replies 2 ·
Replies
2
Views
4K