Explain the practical strategy for using the Poincare-Bendixson theorem to prove a limit cycle exists in a specific system.
Think about your answer, then reveal below.
Model answer: Construct a trapping region — a bounded, closed subset of the phase plane that trajectories enter but cannot leave (the vector field points inward on the boundary). Then show that the trapping region contains no stable fixed points (either no fixed points at all, or only unstable ones). By Poincare-Bendixson, the omega-limit set of any trajectory in the region must be a periodic orbit. Often the trapping region is an annulus: the outer boundary is established using a Lyapunov-like argument showing trajectories can't escape to infinity, and the inner boundary comes from the instability of a fixed point (trajectories spiral away from it).
For the van der Pol oscillator, the strategy works beautifully. Near the origin, the fixed point is unstable (eigenvalues have positive real part for μ > 0), so trajectories spiral outward — the inner boundary of the annulus. Far from the origin, the strong damping (x² ≫ 1) drives trajectories inward, establishing the outer boundary. The annulus contains no fixed points (the only one is at the origin, excluded by the inner boundary). Poincare-Bendixson guarantees a limit cycle exists between the boundaries.