Solve second order diff equation using substitution

Click For Summary

Homework Help Overview

The problem involves solving a second-order differential equation through substitution. The original poster presents the equation d²y/dx² - dy/dx + y*exp(2x) = x*exp(2x) - 1 and suggests substituting t = exp(x) and setting z(t) = y(x) to rewrite the equation.

Discussion Character

  • Exploratory, Conceptual clarification, Mathematical reasoning

Approaches and Questions Raised

  • Participants discuss the proper application of the chain rule in changing variables and derivatives. There is an emphasis on correctly transforming the derivatives when substituting variables. Some participants suggest using variation of parameters for finding particular solutions.

Discussion Status

The discussion is ongoing, with participants providing guidance on the correct differentiation process and exploring different methods for solving the equation. There is no explicit consensus on the approach yet, but several lines of reasoning are being explored.

Contextual Notes

Participants note the importance of correctly applying the chain rule and the implications of variable substitution on the derivatives. There is also mention of the related homogeneous equation and its solutions, indicating that the problem is part of a broader context of differential equations.

Gekko
Messages
69
Reaction score
0

Homework Statement



d2y/dx2-dy/dx+y*exp(2x) = x*exp(2x)-1

substitute t=exp(x) and set z(t)=y(x) and rewrite hence find all solutions

The Attempt at a Solution



Rewriting gives:
d2z/dt2-dz/dt+z*t^2=(ln(t) * t^2) - 1

however I don't see how this in any way helps us...
 
Physics news on Phys.org
You haven't changed the derivatives properly. You cannot just replace "dy/dx" with "dy/dt". You have to use the chain rule.

dy/dx= (dy/dt)(dt/dx) so if t= e^x, dy/dx= e^x(dy/dt)= t(dy/dt).

d^2y/dx^2= (d/dx)(dy/dx)= d/dx(t dy/dt)= t(d/dt(t(dy/dt)). By the product rule, that is t^2 d^2y/dt^2+ t dy/dt.
The equation becomes (t^2 d^2y/dt^2+ t dy/dt)- t dy/dt+ t^2 y= t^2(d^2y/dt^2+ y)= t^2ln(t)- 1 so that d^2y/dt^2+ y= ln(t)- t^{-2}.

The related homogeneous equation is easy to solve and I would recommend "variation of parameters" to find a specific solution to the entire equation.
 
first of all:
"d2y/dx2" does not equals "d2z/dt2" and "dy/dx" does not equals "dz/dt".

but you right in saying that exp(2x) = t^2 and x = ln(t)

you need to use the "chain rule" in differentiation !

t=e^x

z(t)=y(x)

dz(t)/dx = dy(x)/dx => (dz(t)/dt)*(dt/dx) = y'(x)

*try to go from here :)
*don't forget the differentiate correctly when you do d^2z(t)/dx^2 = d^2y(x)/dx^2, it has the form (u*v)' = u'*v + u*v'

Hope it will lead you to the right answer ! :)
 
HallsofIvy beat me to it :)

also I love Lagrange's variation of parameters !

Here is how you do it:

(( in your case y=z, x=t ))

y1(x), y2(x) : are two linearly independent homogeneous solutions.
g(x) : the particular expression, in your case = ln(t)- t^{-2}.

u'1(x)*y1(x) + u'2(x)*y2(x) = 0
u'1'(x)*y'1(x) + u'2(x)*y'2(x) = g(x)

solve the system of equations and find u1(x) and u2(x) by integration.
the particular part of the answer will be:

Y(x) = u1(x)*y1(x) + u2(x)*y2(x).

the complete answer for the ODE (A,B are constants):

y(x) = [A*y1(x)] + [B*y1(x)] + [u1(x)*y1(x) + u2(x)*y2(x)].

*Lagrange was a genius !
 

Similar threads

Replies
4
Views
2K
  • · Replies 3 ·
Replies
3
Views
2K
Replies
1
Views
2K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 5 ·
Replies
5
Views
2K
  • · Replies 105 ·
4
Replies
105
Views
11K
  • · Replies 2 ·
Replies
2
Views
1K
  • · Replies 8 ·
Replies
8
Views
2K
  • · Replies 2 ·
Replies
2
Views
1K
  • · Replies 1 ·
Replies
1
Views
2K