ProofComplete

Two-Dimensional Flows - Key Proof

ProofProof Sketch of Poincare-Bendixson Theorem

We provide a conceptual outline of the proof of the Poincare-Bendixson theorem, focusing on the key ideas rather than technical details.

Theorem Statement (Simplified Version): Let γ\gamma be a non-closed trajectory in R2\mathbb{R}^2 that remains in a closed, bounded region KK containing no fixed points for all t0t \geq 0. Then the ω\omega-limit set of γ\gamma is a closed orbit.

Proof Outline:

Step 1: Properties of ω\omega-limit sets. The ω\omega-limit set ω(γ)\omega(\gamma) is the set of accumulation points of {ϕt(x0):t0}\{\phi_t(\mathbf{x}_0) : t \geq 0\}. Since KK is compact and contains the forward trajectory, ω(γ)\omega(\gamma) is nonempty, compact, and invariant under the flow. By assumption, ω(γ)\omega(\gamma) contains no fixed points.

Step 2: Non-crossing of trajectories. In planar systems, distinct trajectories cannot cross (by uniqueness of solutions to ODEs). This topological constraint is crucial. If two trajectories crossed at a point pp, there would be two different solutions starting from pp, contradicting uniqueness.

Step 3: Poincare section and return map. Choose a point pω(γ)p \in \omega(\gamma) and construct a small transverse section Σ\Sigma through pp—a line segment perpendicular to the vector field. Since pp is in the ω\omega-limit set, the trajectory γ\gamma returns arbitrarily close to pp infinitely often, hence intersects Σ\Sigma infinitely many times.

Let {pn}\{p_n\} be the sequence of intersection points. Since Σ\Sigma is one-dimensional and bounded, and trajectories cannot cross, the sequence {pn}\{p_n\} must be monotonic along Σ\Sigma (always approaching from the same side). Therefore, {pn}\{p_n\} converges to some point qΣq \in \Sigma.

Step 4: The limit point qq lies in ω(γ)\omega(\gamma). By construction, q=limnpnq = \lim_{n \to \infty} p_n is an accumulation point of the trajectory, so qω(γ)q \in \omega(\gamma). Moreover, since ω(γ)\omega(\gamma) is invariant, the entire trajectory through qq lies in ω(γ)\omega(\gamma).

Step 5: The trajectory through qq is closed. Consider the trajectory γq\gamma_q starting at qq. This trajectory must return to Σ\Sigma again at some point qq' (since it stays in the compact region ω(γ)\omega(\gamma) and cannot approach a fixed point). By invariance of ω(γ)\omega(\gamma), we have qω(γ)q' \in \omega(\gamma).

The key insight: if qqq' \neq q, then by the monotonicity argument in Step 3, the sequence would continue moving along Σ\Sigma, contradicting the fact that qq is the limit. Therefore, q=qq' = q, meaning the trajectory returns exactly to its starting point. Thus, γq\gamma_q is a closed orbit.

Step 6: The original trajectory approaches the closed orbit. Having established that ω(γ)\omega(\gamma) contains a closed orbit γq\gamma_q, we must show ω(γ)=γq\omega(\gamma) = \gamma_q (i.e., the limit set consists of exactly this orbit). This follows from analyzing the Poincare map on Σ\Sigma: if ω(γ)\omega(\gamma) contained points not on γq\gamma_q, the non-crossing property and compactness would be violated.

Conclusion: Any trajectory confined to a compact region without fixed points must approach a closed orbit. Combined with the case where fixed points are present (in which case trajectories approach fixed points or heteroclinic connections), this gives the complete Poincare-Bendixson theorem.

This proof illustrates several powerful techniques in dynamical systems: reduction to Poincare sections (lowering dimension), exploitation of topological constraints (non-crossing), and compactness arguments. The restriction to two dimensions is essential; the non-crossing property fails in three or more dimensions, where trajectories have "room" to weave around each other, enabling chaotic behavior.

Remark

The Poincare-Bendixson theorem's proof relies fundamentally on planar topology. In three dimensions, two curves can pass by each other without intersecting, so the monotonicity argument fails. This is why the theorem doesn't extend to higher dimensions, and why three-dimensional flows can exhibit chaos. The Lorenz attractor, for instance, has a bounded trajectory that approaches neither a fixed point nor a periodic orbit but instead follows a fractal strange attractor—behavior impossible in two dimensions.

ExampleApplication to Van der Pol Equation

For the Van der Pol oscillator x¨μ(1x2)x˙+x=0\ddot{x} - \mu(1-x^2)\dot{x} + x = 0 with μ>0\mu > 0, we can apply Poincare-Bendixson:

  1. The origin is an unstable fixed point (source)
  2. For large x|x| or x˙|\dot{x}|, trajectories point inward (can be shown by considering V=x2+x˙2V = x^2 + \dot{x}^2)
  3. Therefore, there exists an annular region KK bounded away from the origin where trajectories enter but don't leave
  4. By Poincare-Bendixson, KK contains a closed orbit

This proves the existence of a limit cycle without computing it explicitly—a typical application of the theorem.

The proof of Poincare-Bendixson exemplifies how topological reasoning complements analytical techniques in dynamical systems. Without solving differential equations, we derive powerful qualitative conclusions about asymptotic behavior. This geometric viewpoint, pioneered by Poincare, revolutionized the study of differential equations and remains central to modern dynamical systems theory.