Discrete Dynamical Systems - Key Properties

Discrete maps possess structural properties governing their dynamics. Invertibility, ergodicity, and mixing determine how maps redistribute points over time, while normal forms simplify analysis near bifurcations by eliminating inessential nonlinear terms.

DefinitionInvertibility and Noninvertibility

A map $f: X \to X$ is invertible if there exists a map $f^{-1}: X \to X$ such that $f \circ f^{-1} = f^{-1} \circ f = \text{id}$ . Invertible maps preserve information: knowing $x_n$ determines $x_{n-1}$ uniquely. Noninvertible maps can have multiple preimages: $f^{-1}(\{y\})$ may contain several points.

Conservative maps (area-preserving, symplectic) are typically invertible
Dissipative maps (contracting volumes) are often noninvertible
Noninvertibility creates fractal basin boundaries and can lead to chaotic attractors

Invertibility profoundly affects dynamics. Invertible maps like the standard map exhibit conservative chaos where phase space volume is preserved. Noninvertible maps like the logistic map exhibit dissipative chaos with attractors of measure zero. The Henon map, with small Jacobian $|\det Df| = 0.3 < 1$ , is weakly dissipative yet invertible, bridging these extremes.

DefinitionInvariant Measures

A measure $\mu$ on $X$ is invariant under $f$ if $\mu(f^{-1}(A)) = \mu(A)$ for all measurable sets $A$ . Equivalently, for any integrable function $\phi$ :

$\int_X \phi(f(x)) \, d\mu(x) = \int_X \phi(x) \, d\mu(x)$

Invariant measures describe stationary statistical properties of the dynamics. For ergodic systems, time averages equal space averages almost everywhere.

Invariant measures generalize the notion of equilibrium to chaotic systems. While individual trajectories are unpredictable, their statistical distribution may be stationary. The natural measure (SRB measure) is often physically relevant, describing long-term statistics observed in simulations and experiments.

DefinitionUnimodal Maps

A map $f: [a,b] \to [a,b]$ is unimodal if it has a single local maximum (critical point) $c$ where $f'(c) = 0$ . The behavior of the critical point's orbit determines global dynamics:

If the critical orbit approaches a stable periodic orbit, almost all orbits do likewise
If the critical orbit is dense, the map is chaotic
The kneading sequence (symbolic itinerary of $c$ ) completely characterizes topological conjugacy class

Unimodal maps include the logistic family and are paradigmatic examples for understanding one-dimensional dynamics.

ExampleStandard Map

The standard (Chirikov) map models periodically kicked oscillators:

$x_{n+1} = x_n + y_n, \quad y_{n+1} = y_n + K \sin(x_n)$

For small $K$ , invariant circles persist (KAM theory). As $K$ increases:

KAM curves break, creating chaotic zones
At $K \approx 0.97$ , the last KAM curve (golden mean rotation) breaks
For large $K$ , global chaos with diffusive transport

The standard map demonstrates the transition from integrable to chaotic dynamics as perturbation strength increases.

Remark

Properties like invertibility, ergodicity, and the structure of critical point orbits determine whether discrete systems behave regularly or chaotically. Unimodal maps, despite their simplicity, exhibit the full complexity of chaotic dynamics and serve as testing grounds for general theories. Understanding these properties allows classification of maps and prediction of qualitative behavior across parameter space without exhaustive numerical exploration.

These properties reveal that discrete dynamics, while superficially simpler than continuous flows, contains equal depth and complexity. The interplay between algebraic (invertibility), geometric (critical points), and measure-theoretic (invariant measures) perspectives provides a complete picture of how discrete maps evolve systems over time.