2
$\begingroup$

I strive to prove that the following system of differential equations:

$$\begin{cases} x'=x-u(t)xy\\ y'= -y+u(t)xy \\ x(0)=x_0>0\\ y(0)=y_0>0 \end{cases}$$

has a unique Caratheodory solution on a given interval $[0,T]$, where $u:[0,T]\to [0,1]$ is a control, lets say measurable or continuous if necessary. I cannot apply Caratheodory existence theorem https://en.wikipedia.org/wiki/Carath%C3%A9odory%27s_existence_theorem since the function $f:[0,T]\times\mathbb{R}^2\to\mathbb{R}^2,\ f(t,x,y)=(x-u(t)xy,-y+u(t)xy)$ is not satisfying the Lipschitz-like condition, and the linearity-like growing in $(x,y)$. Someone told me that maybe it can be proven that the solution $(x,y)$ if it exists is bounded (via Lyapunov function or via first inegral), but I cannot find a solution yet.

$\endgroup$
1
  • $\begingroup$ $x$ and $y$ cannot cross into the negative area without the other variable getting unbounded. On the other hand, as long as $x,y>0$, the sum $x+y$ can grow at most exponentially in $t$. $\endgroup$
    – fedja
    Nov 10, 2018 at 2:11

1 Answer 1

3
$\begingroup$

Assuming that $u$ is Lebesgue integrable, $f$ does satisfy a Lipschitz-like condition, so we have (local) existence and uniqueness theorem.

Whatever the controls, the sets $\{(0,0)\}$, $\{\, (x, 0): x > 0\,\}$ and $\{\, (0, y): y > 0\,\}$ are invariant. So, by uniqueness, any nonextendible solution starting in $\mathbb{R}_{++} := \{\, (x , y): x > 0,\ y > 0 \,\}$ stays there (as fedja noticed in their comment).

Suppose to the contrary that for some initial conditions $(x_0, y_0) \in \mathbb{R}_{++}$ the solution blows up at $\tau \in (0, T)$. Since $1 - u(t) y(t) \le 1$, by standard differential inequalities we have that $x(t) \le x_0 e^{\tau}$ for all $t \in [0, \tau)$. Consequently, as $-1 + u(t) x(t) \le -1 + x_0 e^{\tau}$, again by differential inequalities we have $y(t) \le y_0 \exp{((-1 + x_0 e^{\tau}) \tau)}$ for all $t \in [0, \tau)$, a contradiction.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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