I guys. I'm doing this problem: $$\begin{cases} y'(x)=\frac{2y(x)}{x}+3x^{\alpha} \\ y(1)=2\\ \end{cases} $$ with $\alpha\in \mathbb R^{+}$ constant.

I have started studying the existence and uniqueness of solutions. So $$f(x,y)=\frac{2y(x)}{x}+3x^{\alpha}$$ Has domain $D=\{(x,y)\in\mathbb R^{2} :x\neq0\}$.

I notice that $f(x,y$) is a continuous function on it's domain, so it's a derivable function and for this reason I can say that is locally Lipschitz continuous. So I can apply the local exinstance and uniqueness theorem.

Moreover $f(x,y)$ it's not Lipschitz in $y$ respect to $x$ on $D$, so it's not globally Lipschitz continuous and I can't apply the global existance and uniquess theorem for the solutions.

Now I want to find explicitly the solutions, depending on $\alpha$, and finally I want to find the value of $\alpha$ for which

$$\lim_{x\to \infty}y(x)=+\infty$$

Assuming all is correct from the beginning to this point, the equation could be solved by separation of variables:


Is this the most convenient possibility to solve this equation? Because I'm a little bit stucked at this point. Any suggestions to get the goal? Thanl you very much.


The Cauchy problem has a unique global solution in $(0,+\infty)$ because the differential equation is linear.

Moreover, after multiplying by the integrating factor $1/x^2$ we have: $$\left(\frac{y(x)}{x^2}\right)'=\frac{y'(x)}{x^2}-\frac{2y(x)}{x^3}=3x^{\alpha-2}.$$ Then after integrating both side we obtain:

i) if $\alpha\not=1$, $$\frac{y(x)}{x^2}=\int 3x^{\alpha-2}\,dx=\frac{3x^{\alpha-1}}{\alpha-1}+C\implies y(x)=\frac{3x^{\alpha+1}}{\alpha-1}+Cx^2;$$

ii) if $\alpha=1$, $$\frac{y(x)}{x^2}=\int 3x^{-1}\,dx=3\ln(x)+C \implies y(x)=3x^2\ln(x)+Cx^2.$$ Can you take it from here?

  • $\begingroup$ Thank you for your reply. I am thinking about the domain. For sure I have $\alpha>0$. But why $x$ couldn't be negative? If i pick (just for an example) $x=-1$ and $\alpha=2$ I would have $$3(-1)^{2}=9$$ That is a real positive number. What do you think about it? Am I strongly in error? Thank you. Edit: Maybe it gives some trouble when $0\leq\alpha\leq 1$ because we're going into $\mathbb C$. $\endgroup$ – muserock92 Feb 20 '18 at 15:07
  • $\begingroup$ Usually the domain of the solution of a Cauchy problem is an interval (a connected set) which contain the initial point ($1$ in this case). Moreover if $\alpha$ is irrational $x^{\alpha}$ is not defined for $x<0$ (what is $(-2)^{\pi}$?) $\endgroup$ – Robert Z Feb 20 '18 at 15:13
  • $\begingroup$ Yes, It's definitely the point of my edit. We don't want to work with compex numbers. $\endgroup$ – muserock92 Feb 20 '18 at 15:14
  • $\begingroup$ And what about existance and uniqueness of solutions? I just don't care about theorems, I just see that it's linear so for sure I have a unique global solution. Is it ok now? Thank you. $\endgroup$ – muserock92 Feb 20 '18 at 15:15
  • $\begingroup$ Yes! Moreover here we also found explicit solutions which are defined for $x>0$. $\endgroup$ – Robert Z Feb 20 '18 at 15:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.