I'm a little bit confused with the theorem of existence and uniqueness for ODEs (the Cauchy-Lipschitz or Picard-Lindelof theorem if I'm right). Let assume we can apply the theorem on the ODEs $X'=F(t,X)$ in $\mathbb{R}^n$. Let consider that initial conditions are in the positive quadrant $X(0) \in \mathbb{R}_+^n$ (I mean $X(0)>0$.

Is it true to say that if there is the trivial equilibrium $F(t,0^n)=0$, then all solution starting in the positive quadrant $\mathbb{R}_+^n$ stay in $\mathbb{R}_+^n$ ?

The point is that I found a paper in ecology where authors study a system of ODEs on 2 dimensions (in $\mathbb{R}^2$) with $(0,0)$ an equilibrium, and they assume $x_1(0)>0$ and $x_2(0)>0$. Then, they write that theorem of existence and uniqueness ensures positives solutions and the axis cannot intersect and therefore: $x_1(t) >0$ and $x_2(t)$. I don't understand why the trajectory cannot turn around the point $(0,0)$ and so intersect the axis.


The result is false in general. Counterexample: $$ x'=y,\quad y'=-x. $$ Trajectories are circles centered at $0$.

However, it might be true for the specific differential equation modeling the ecological system in the paper. That will depend on the properties of the function $F$. If for instance $$ x'=f(x,y),\quad y'=g(x,y) $$ with $f,g\ge0$, both $x$ and $y$ will be increasing.

  • $\begingroup$ Thank you. I actually missed the information that $f(0,y)=0$ and $g(x,0)=0$, and so axis bound the system. $\endgroup$ – Leon-Alph Dec 4 '14 at 16:34

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.