Dynamical Systems
Bifurcation Theory
Bifurcations are points of qualitative change in a system's behavior under continuous parameter variation. The same mathematics describes a beam buckling under compression, cardiac fibrillation onset, laser threshold crossing, and epidemic threshold - one framework for all.
- Cardiology: transition from sinus rhythm to atrial fibrillation is a Hopf bifurcation at a critical stimulation frequency (Glass 1979, basis for next-generation pacemakers)
- Structural mechanics: Euler's critical buckling load for a beam is a pitchfork bifurcation - the compressed beam snaps to one side when load exceeds the critical value
- Laser physics: lasing threshold is a saddle-node bifurcation as pump power crosses the threshold value
- Neuroscience: transition of a neuron from rest to firing at threshold current is a Hopf bifurcation in the Hodgkin-Huxley model
Предварительные знания
- Phase portraits
- Equilibrium stability
- Linearization and Jacobian
Saddle-node and transcritical bifurcations
In 1972, Rene Thom published catastrophe theory: 7 elementary forms describe all bifurcations of systems with at most 2 control parameters. Leon Glass in 1979 showed that the transition from sinus rhythm to fibrillation is a Hopf bifurcation: at a critical stimulation frequency the cardiac rhythm shifts from periodic to chaotic.
Hartman-Grobman theorem: near a hyperbolic fixed point (Re(lambda_i) != 0) the phase portrait is topologically conjugate to the linearization. This justifies classifying equilibria via the Jacobian J = Df(x*). Hyperbolicity fails at bifurcation points.
Application: laser threshold. As pump P increases from zero, the photon number in the cavity stays at zero (stable). At critical P_c a pair of equilibria appears and one becomes stable - the laser starts emitting. This is a saddle-node bifurcation.
What is the normal form of a saddle-node bifurcation?
Saddle-node normal form: x' = mu - x^2. For mu < 0 no equilibria. At mu = 0 one equilibrium. For mu > 0 two equilibria x* = ±sqrt(mu): stable +sqrt(mu) and unstable -sqrt(mu).
Pitchfork bifurcation and normal forms
The pitchfork bifurcation arises in systems with x → -x symmetry. As the control parameter mu crosses 0, the zero equilibrium loses stability and two symmetric stable states x* = ±sqrt(mu) emerge. The bifurcation diagram looks like a Y - hence the name.
Bifurcation as a branching tree. As parameter mu changes, the system traces a branch of equilibria like a tree trunk. At a bifurcation point the trunk branches: some branches are stable (drawn solid), others unstable (dashed). A pitchfork looks like a Y; a Hopf sends off a closed loop representing the limit cycle.
A normal form is a simplified system topologically equivalent to the original near the bifurcation. Computing normal forms is algorithmic: eliminate non-resonant terms order by order using near-identity transformations.
At which bifurcation do two new stable equilibria emerge from an unstable one?
As mu crosses 0, the zero equilibrium loses stability and two symmetric stable states x* = ±sqrt(mu) are born. This requires the system to have a symmetry x → -x. Example: Eulers buckling of a beam under compression.
Hopf bifurcation and birth of a limit cycle
The Hopf bifurcation differs from the previous ones: instead of new equilibria appearing, a limit cycle is born - an isolated closed orbit. This is the mathematical model of oscillation onset: cardiac rhythm, laser generation, neuronal activity.
Subcritical Hopf bifurcation: for mu < 0 an unstable limit cycle is born (unusual direction), which collapses to a point at mu = 0. This produces hysteresis: the system can jump to a large-amplitude cycle before the linear threshold is reached. This is the mechanism behind cardiac fibrillation and some laser instabilities.
| Bifurcation | Normal form | What happens | Example |
|---|---|---|---|
| Saddle-node | x' = mu - x^2 | Equilibrium pair is born/destroyed | Laser threshold |
| Transcritical | x' = mu*x - x^2 | Equilibria exchange stability | Population growth |
| Pitchfork (super) | x' = mu*x - x^3 | Unstable zero births two stable | Euler beam |
| Hopf (super) | polar: r' = mu*r - r^3 | Equilibrium births limit cycle | Fibrillation |
What is the condition for a Hopf bifurcation?
A Hopf bifurcation requires: 1) at mu=mu_c a conjugate pair lambda = ±i*omega with omega != 0; 2) transversality: d/dmu Re(lambda) != 0 at mu=mu_c. The born cycle has frequency omega and radius ~sqrt(|mu - mu_c|).
Connections to other areas
Bifurcation theory runs through all of nonlinear dynamics and has direct applications in physics, biology, and engineering.
- Catastrophe theory — Related topic
- Neural networks — Related topic
- Control theory — Related topic
- Kuramoto synchronization — Related topic
Итоги
- Saddle-node: normal form x' = mu - x^2; equilibrium pair appears for mu > 0 and disappears for mu < 0
- Transcritical: x' = mu*x - x^2; two equilibria always exist, exchange stability at mu=0
- Pitchfork (supercritical): x' = mu*x - x^3; zero loses stability at mu=0, two stable branches x* = +/-sqrt(mu) appear
- Hopf: conjugate eigenvalue pair crosses imaginary axis, limit cycle of radius ~sqrt(mu) is born
- Hartman-Grobman: near a hyperbolic fixed point, topology of flow determined by Jacobian
- Normal forms: algorithmic simplification near bifurcation by eliminating non-resonant terms
What is the condition for a Hopf bifurcation?
A Hopf bifurcation requires: 1) at mu=mu_c a conjugate pair lambda = +/-i*omega with omega != 0; 2) transversality: d/dmu Re(lambda) != 0 at mu=mu_c. This distinguishes it from degenerate cases. The born cycle has frequency omega and radius ~sqrt(|mu - mu_c|).