Nonlinear Systems

Predator-prey, van der Pol, and limit cycles

Nonlinear Systems

Most real-world differential equations are nonlinear. Unlike linear systems, nonlinear systems can exhibit dramatically different behavior: closed orbits that are not ellipses, limit cycles that attract or repel nearby trajectories, and sensitive dependence on parameters. In this section, we explore three classic nonlinear systems that reveal the richness of nonlinear dynamics.

A central result is the Poincare-Bendixson theorem: in two-dimensional autonomous systems, the only possible long-term behaviors are convergence to an equilibrium, convergence to a limit cycle, or escape to infinity. There is no chaos in the plane -- that requires at least three dimensions.

Lotka-Volterra Predator-Prey Model

The Lotka-Volterra equations model the interaction between prey (x) and predators (y). Prey grow exponentially in the absence of predators, while predators decline without prey. Their interaction creates closed orbits in the phase plane -- perpetual oscillations where prey and predator populations rise and fall in a delayed cycle.

dx/dt = alpha * x - beta * x * y   (prey growth minus predation)
dy/dt = delta * x * y - gamma * y   (predator growth minus death)
Phase Portrait (click to set initial conditions)
Time Series

Key insight: This is a conservative system -- it has a conserved quantity (the Lotka-Volterra integral), so trajectories form closed orbits. The equilibrium point at (gamma/delta, alpha/beta) is a center, not a spiral. Energy is neither added nor dissipated, so oscillations persist forever without damping or growing. Click different starting points to see orbits of different amplitudes.

Van der Pol Oscillator

The Van der Pol oscillator is a dissipative system with a remarkable property: it has a stable limit cycle. Trajectories starting inside the limit cycle spiral outward, and trajectories starting outside spiral inward -- all converging to the same periodic orbit. The parameter mu controls the strength of the nonlinear damping.

x'' - mu(1 - x^2)x' + x = 0
System form: dx/dt = y,   dy/dt = mu(1 - x^2)y - x
Phase Portrait — trajectories converge to the limit cycle
Time Series x(t) — relaxation oscillations for large mu
0.1 (nearly sinusoidal)5.0 (relaxation)

Key insight: For small mu, oscillations are nearly sinusoidal. As mu increases, the system exhibits relaxation oscillations -- long periods of slow change punctuated by rapid jumps. The limit cycle exists for all mu > 0 and is guaranteed by the Poincare-Bendixson theorem: the origin is an unstable equilibrium, and trajectories are bounded, so they must converge to a periodic orbit.

Nonlinear Pendulum Phase Portrait

The simple pendulum is one of the most important nonlinear systems. Its phase portrait reveals three qualitatively different types of motion: libration (back-and-forth swinging), rotation (going over the top), and the separatrix -- the critical energy level that divides the two regimes.

theta'' + (g/L) sin(theta) = 0
System form: d(theta)/dt = omega,   d(omega)/dt = -(g/L) sin(theta)
Pendulum Animation
Phase Portrait (click to set initial condition)
Small oscillationSeparatrix ~ 4.4Full rotation

Key insight: The pendulum is a conservative system (Hamiltonian), so its phase portrait consists of level curves of the energy function E = (1/2)omega^2 - (g/L)cos(theta). The saddle points at theta = +-pi correspond to the unstable upright equilibrium. The separatrix connects these saddle points and represents the trajectory where the pendulum just barely reaches the top with zero velocity.

Conservative vs. Dissipative Systems

Conservative Systems

  • Preserve a conserved quantity (energy, integral of motion)
  • Phase portraits show closed orbits (not limit cycles)
  • Orbits depend on initial conditions -- different ICs give different orbits
  • Examples: pendulum, Lotka-Volterra, Hamiltonian systems

Dissipative Systems

  • Energy is added or removed by nonlinear mechanisms
  • Can have limit cycles -- isolated periodic orbits that attract neighbors
  • Long-term behavior is independent of initial conditions (within basin)
  • Examples: Van der Pol, biological oscillators, electronic circuits

Key Takeaways

  • Limit cycles are isolated periodic orbits unique to nonlinear systems. They cannot occur in linear systems, where periodic orbits always come in continuous families.
  • The Poincare-Bendixson theorem constrains planar dynamics: bounded trajectories that do not approach an equilibrium must approach a limit cycle. This is why chaos requires at least three dimensions.
  • Conservative systems (like the pendulum and Lotka-Volterra) preserve energy, producing families of closed orbits. Dissipative systems(like the Van der Pol oscillator) can have attracting limit cycles.
  • Separatrices are special trajectories that divide the phase plane into regions of qualitatively different behavior, as seen in the pendulum phase portrait.