1. Not finding help here? Sign up for a free 30min tutor trial with Chegg Tutors
    Dismiss Notice
Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Nonlinear differential equation (quick)

  1. Jun 19, 2005 #1
    I need some help with the problem that follows. Any help is highly appreciated.

    Problem:

    "Consider the initial value problem [tex]y^{\prime}=y^{1/3}, \mbox{ }y(0)=0[/tex] from Example 3 in the text.
    (a) Is there a solution that passes through the point (1,1)? If so, find it.
    (b) Is there a solution that passes through the point (2,1)? If so, find it.
    (c) Consider all possible solutions of the given initial value problem. Determine the set of values that these solutions have at [tex]t=2[/tex]."

    Example 3:

    [tex]y^{\prime}=y^{1/3}, \qquad y(0)=0 \quad \qquad (1)[/tex]

    [tex]y^{-1/3} \: dy = dt[/tex]

    [tex]\int y^{-1/3} \: dy = \int dt[/tex]

    [tex]\frac{3}{2} y^{2/3} = t + \mathrm{C}[/tex]

    [tex]y = \left[ \frac{2}{3} \left( t + \mathrm{C} \right) \right] ^{3/2}[/tex]

    The initial condition is satisfied if [tex]\mathrm{C}=0[/tex], so

    [tex]y=\phi _1 (t) = \left( \frac{2}{3}t \right) ^{3/2}, \qquad t \geq 0 \quad \qquad (2)[/tex]

    On the other hand, the function

    [tex]y=\phi _2 (t) = -\left( \frac{2}{3}t \right) ^{3/2}, \qquad t \geq 0 \quad \qquad (3)[/tex]

    is yet another solution. For an arbitrary positive [tex]t_0[/tex] the functions

    [tex]y=\left\{ \begin{array}{ll} 0, &\mbox{ if } 0 \leq t < t_0 \\ \pm \left[ \frac{2}{3} \left( t - t_0 \right) \right] ^{3/2} , &\mbox{ if } t \geq t_0 \end{array} \right. \quad \qquad (4)[/tex]

    are solutions of Eq. (1).

    Answers:

    "(a) No. (b) Yes; set [tex]t_0=\frac{1}{2}[/tex] in Eq. (4) from Example 3 (c) [tex]\left| y \right| \leq \left( \frac{4}{3} \right) ^{3/2}[/tex]"

    My work:

    (a) (Why is this wrong?)

    [tex]y^{\prime}=y^{1/3}, \qquad y\left( 1 \right)=1[/tex]

    [tex]y^{-1/3} \: dy = dt[/tex]

    [tex]\int y^{-1/3} \: dy = \int dt[/tex]

    [tex]\frac{3}{2} y^{2/3} = t + \mathrm{C}[/tex]

    [tex]y = \left[ \frac{2}{3} \left( t + \mathrm{C} \right) \right] ^{3/2}[/tex]

    The initial condition is satisfied if [tex]\mathrm{C}=\frac{1}{2}[/tex], so

    [tex]y= \left[ \frac{2}{3} \left( t + \frac{1}{2} \right) \right] ^{3/2}[/tex]

    which seems to work.

    (b) (This one seems to be ok.)

    [tex]y^{\prime}=y^{1/3}, \qquad y\left( 2 \right)=1[/tex]

    [tex]y^{-1/3} \: dy = dt[/tex]

    [tex]\int y^{-1/3} \: dy = \int dt[/tex]

    [tex]\frac{3}{2} y^{2/3} = t + \mathrm{C}[/tex]

    [tex]y = \left[ \frac{2}{3} \left( t + \mathrm{C} \right) \right] ^{3/2}[/tex]

    The initial condition is satisfied if [tex]\mathrm{C}=-\frac{1}{2}[/tex], so

    [tex]y= \left[ \frac{2}{3} \left( t - \frac{1}{2} \right) \right] ^{3/2}[/tex]

    (c) (This one seems to be ok.)

    Considering all possible solutions of the given initial value problem, the maximum value at [tex]t=2[/tex] is given by the solution

    [tex]y = \left( \frac{2}{3}t \right) ^{3/2} \Rightarrow y(2) = \left( \frac{4}{3} \right) ^{3/2}[/tex]
     
  2. jcsd
  3. Jun 19, 2005 #2

    StatusX

    User Avatar
    Homework Helper

    I'm a little confused. Is the y(0)=0 condition applied through the whole problem? If so, there is one unique solution, since this is a first order differential equation. Unless you're supposed to consider the cube root as a multivalued function, in which case the other solutions will be complex.
     
  4. Jun 19, 2005 #3

    saltydog

    User Avatar
    Science Advisor
    Homework Helper

    I don't believe the Lipschitz condition is met with:

    [tex]\frac{\partial}{\partial y} y^{1/3}[/tex]

    That is why we obtain two solutions passing through (0,0), that is:

    [tex]y(x)=[\pm\sqrt{(2/3)x}]^3[/tex]

    Edit: I deleted the earlier post I made. Misinterpreted the problem. Just easier to remove it.

    Edit: Also, y(x)=0 is a solution.
     
    Last edited: Jun 19, 2005
  5. Jun 19, 2005 #4

    saltydog

    User Avatar
    Science Advisor
    Homework Helper

    One more Thiago:

    For the initial value equation:

    [tex]y^{'}=y^{1/3};\qquad y(0)=0[/tex]

    We find three solutions:

    [tex]y(x)=0[/tex]

    [tex]y(x)=[\pm\sqrt{(2/3)x}]^3[/tex]

    So, for a and b I say no.
     
  6. Jun 19, 2005 #5

    StatusX

    User Avatar
    Homework Helper

    Those solutions do appear to work. It seems strange though, because if you start at (0,0), the initial slope is 0, so y doesn't change, and so the slope stays 0. I don't understand how the function starts to curve away from the x-axis.
     
  7. Jun 20, 2005 #6
    That annoyed me at first too. If the initial condition were different for parts (a) and (b), then it'd be so nice! I assumed it worked this way, but I am wrong. In fact, it seems that I should focus on Eq. (4), in particular:

    [tex]y = \frac{2}{3} \left( t - t_0 \right) \right] ^{3/2}[/tex]

    Frankly, I don't see how [tex]t_0[/tex] comes into play, and why [tex]t_0[/tex] should equal [tex]\frac{1}{2}[/tex] for part (b). Could you please clarify that?

    I think that's because t increases, which implies that y increases.

    Thank you so much.
     
  8. Jun 20, 2005 #7

    OlderDan

    User Avatar
    Science Advisor
    Homework Helper

    Lot's of functions have zero slope somewhere, and all power functions with exponent > 1 have zero slope at (0,0), but they change. With a properly chosen coefficient, the derivative of these power functions is a root of the function.

    [tex] y = \left(\frac {t}{n}\right)^n = \frac {t^n}{n^n} \ \rm{,} \ n>1 \ \rm{,} \ t \ge 0 [/tex]

    [tex] y' = n\frac {t^{n-1}}{n^n} = \frac{n}{t}y = \frac{ny}{ny^{1/n}} = y^{(n-1)/n} [/tex]

    One of the solutions to this problem is the case n = 3/2
     
  9. Jun 20, 2005 #8

    saltydog

    User Avatar
    Science Advisor
    Homework Helper

    Well I have some questions:

    1. How come we're getting 3 solutions for this particular initial-value problem? Why not just two or a handfull or even an infinite number like for this one:

    [tex]ty^{''}+(t-1)y^{'}+y=t^2;\qquad y(0)=0,\quad y^{'}(0)=0[/tex]

    2. What determines how many solutions we get?

    3. Just how do you pronounce "Lipschitz" anyway?

    4. What is the connection between the Lipschitz condition and the number of solutions?

    5. How about a plot?

    Alright, one more:

    6. Can we be guaranteed a unique solution in any neighborhood not including the origin?
     
    Last edited: Jun 20, 2005
  10. Jun 20, 2005 #9

    saltydog

    User Avatar
    Science Advisor
    Homework Helper

    This one too has an infinite number of solutions:

    [tex]y^{'}=\left\{
    \begin{array}{rc1}
    \frac{4x^3y}{x^4+y^2} &\mbox{for} & x,y\neq 0 \\
    0 &\mbox{for} & x,y=0
    \end{array}\right
    [/tex]

    satisfying the initial conditions:

    [tex]y(0)=0[/tex]
     
    Last edited: Jun 20, 2005
  11. Jun 20, 2005 #10
    Folks, I finally figured it out. Thank you very much!

    This is what "Example 3" SHOULD be:

    [tex]y^{\prime}=y^{1/3}, \qquad y\left( t_0 \right)= 0[/tex]

    [tex]y^{-1/3} \: dy = dt[/tex]

    [tex]\int y^{-1/3} \: dy = \int dt[/tex]

    [tex]\frac{3}{2} y^{2/3} = t + \mathrm{C}[/tex]

    [tex]y = \left[ \frac{2}{3} \left( t + \mathrm{C} \right) \right] ^{3/2}[/tex]

    The initial condition is satisfied if [tex]\mathrm{C}=-t_0[/tex], so

    [tex]y = \eta _1 = \left[ \frac{2}{3} \left( t - t_0 \right) \right] ^{3/2}, \qquad t \geq t_0[/tex]

    On the other hand, the function

    [tex]y = \eta _2 = - \left[ \frac{2}{3} \left( t - t_0 \right) \right] ^{3/2}, \qquad t \geq t_0[/tex]

    is yet another solution. Furthermore, the function

    [tex]y = \eta _3 = 0, \qquad 0 \leq t < t_0[/tex]

    also is a solution.

    (a) We are given the point (1, 1). Then, we use [tex]\eta _1[/tex]:

    [tex]y = \left[ \frac{2}{3} \left( t - t_0 \right) \right] ^{3/2}[/tex]

    [tex]1 = \left[ \frac{2}{3} \left( 1 - t_0 \right) \right] ^{3/2} \Rightarrow t_0 = -\frac{1}{2}[/tex]

    Thus, [tex]t_0[/tex] lies outside the domain, which implies that the solution does not pass through the point (1, 1).

    (b) We are given the point (2, 1). Then, we use [tex]\eta _1[/tex]:

    [tex]y = \left[ \frac{2}{3} \left( t - t_0 \right) \right] ^{3/2}[/tex]

    [tex]1 = \left[ \frac{2}{3} \left( 2 - t_0 \right) \right] ^{3/2} \Rightarrow t_0 = \frac{1}{2}[/tex]

    Thus, [tex]t_0[/tex] lies within the domain, which implies that the solution passes through the point (2, 1). We have

    [tex]y = \left[ \frac{2}{3} \left( t - \frac{1}{2} \right) \right] ^{3/2}[/tex]

    (c) Post #1 (seems to be ok).
     
    Last edited: Jun 20, 2005
  12. Jun 20, 2005 #11

    saltydog

    User Avatar
    Science Advisor
    Homework Helper

    Thiago . . . I don't want to be a pain really because I enjoy working the problems you propose but for the record, I disagree with (a). As I see it,

    [tex]y(t) = \left[ \frac{2}{3} \left( t+1/2 \right) \right] ^{3/2} [/tex]

    Is a solution satisfying:

    [tex]y^{'}=y^{1/3}\qquad y(-1/2)=0 \quad\text{and}\quad y(1)=1[/tex]

    As I understand it, the solution y(t)=0 is only when t<-1/2.
     
  13. Jun 20, 2005 #12
    Consider the following

    [tex]y = \left[ \frac{2}{3} \left( t - t_0 \right) \right] ^{3/2}, \qquad t \geq t_0 \qquad (1)[/tex]

    [tex]y = 0, \qquad 0 \leq t < t_0 \qquad (2)[/tex]

    As you can see, each solution has a different domain, but collectively we can say that the domain consists of nonnegative real numbers. In part (a) of my previous post, I let [tex]t=1[/tex] and [tex]y=1[/tex]. Solving Eq. (1) for [tex]t_0[/tex] gives [tex]t_0 = -\frac{1}{2}[/tex]. As a result, it invalidates the statement. Then, it follows that the solution does not pass through the point (1, 1). It seems to me that we have the same thing.

    I appreciate your input.
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?