ProofComplete

One-Dimensional Dynamics - Key Proof

ProofProof of Linearization Near Hyperbolic Fixed Points

We prove that near a hyperbolic fixed point, the dynamics of a smooth map is conjugate to its linearization. This fundamental result justifies the stability analysis via derivatives.

Theorem: Let f:RRf: \mathbb{R} \to \mathbb{R} be C1C^1 with a fixed point at x=0x^* = 0 (by translation). If f(0)<1|f'(0)| < 1 (attracting case), then there exists a neighborhood UU of 00 and a homeomorphism h:Uh(U)h: U \to h(U) such that:

h(f(x))=f(0)h(x)h(f(x)) = f'(0) \cdot h(x)

for all xUx \in U with f(x)Uf(x) \in U. That is, ff is locally conjugate to the linear map xλxx \mapsto \lambda x where λ=f(0)\lambda = f'(0).

Proof:

Step 1: Setup and normalization. Write f(x)=λx+g(x)f(x) = \lambda x + g(x) where λ=f(0)\lambda = f'(0) and g(x)=o(x)g(x) = o(x) as x0x \to 0. We seek hh such that hf=λhh \circ f = \lambda \cdot h near 0, with h(0)=0h(0) = 0 and h(0)=1h'(0) = 1.

Step 2: Formal construction. Assume hh exists with power series h(x)=x+a2x2+a3x3+h(x) = x + a_2 x^2 + a_3 x^3 + \cdots. The conjugacy equation h(f(x))=λh(x)h(f(x)) = \lambda h(x) yields:

h(λx+g(x))=λ(x+a2x2+a3x3+)h(\lambda x + g(x)) = \lambda(x + a_2 x^2 + a_3 x^3 + \cdots)

Expanding the left side using Taylor series and comparing coefficients determines the aka_k uniquely. For example, if g(x)=bx2+O(x3)g(x) = bx^2 + O(x^3), the x2x^2 coefficient gives:

λa2+b=λa2\lambda a_2 + b = \lambda a_2

which forces b=0b = 0 if λ1\lambda \neq 1. More generally, the condition λ<1|\lambda| < 1 ensures the series converges in a neighborhood of 0.

Step 3: Rigorous construction via graph transform. Define the map TT on functions ϕ:UR\phi: U \to \mathbb{R} with ϕ(0)=0\phi(0) = 0 by:

(Tϕ)(x)=1λϕ(f(x))(T\phi)(x) = \frac{1}{\lambda}\phi(f(x))

Under the condition λ<1|\lambda| < 1, the operator TT is a contraction on the space of Lipschitz functions with appropriate norm. By the Banach fixed point theorem, TT has a unique fixed point hh satisfying:

h(x)=1λh(f(x))h(x) = \frac{1}{\lambda}h(f(x))

which rearranges to h(f(x))=λh(x)h(f(x)) = \lambda h(x).

Step 4: Verification of homeomorphism property. The function hh constructed above satisfies h(0)=0h(0) = 0 and h(0)=1h'(0) = 1 by the contraction mapping construction. For xx near 0, we have h(x)1<ϵ|h'(x) - 1| < \epsilon for small ϵ\epsilon, so hh is a local diffeomorphism and hence a homeomorphism on a sufficiently small neighborhood.

Step 5: Extension to repelling case. For f(0)>1|f'(0)| > 1 (repelling case), we apply the above argument to f1f^{-1} in a neighborhood where the inverse exists. The map f1f^{-1} has an attracting fixed point at 0 with derivative (f(0))1(f'(0))^{-1} satisfying (f(0))1<1|(f'(0))^{-1}| < 1. Thus f1f^{-1} is linearizable, and consequently ff is also linearizable.

This linearization theorem, often called the Grobman-Hartman theorem in the one-dimensional case, is fundamental to local bifurcation theory. It guarantees that near hyperbolic fixed points (where f(x)1|f'(x^*)| \neq 1), the nonlinear dynamics is qualitatively the same as the linear approximation.

Remark

The proof fails precisely when f(0)=1|f'(0)| = 1, the neutral case. Here, the linearization provides no information about stability, and higher-order terms in the Taylor expansion become crucial. For example, if f(x)=x+x2f(x) = x + x^2, the fixed point at x=0x = 0 is attracting from one side but repelling from the other, a phenomenon invisible to the linear approximation. This is why bifurcations typically occur at neutral fixed points.

ExampleApplication to Logistic Map

Consider the logistic map fμ(x)=μx(1x)f_\mu(x) = \mu x(1-x) with fixed point x=11/μx^* = 1 - 1/\mu for μ>1\mu > 1. We have:

fμ(x)=2μf'_\mu(x^*) = 2 - \mu

For 1<μ<31 < \mu < 3, we have 2μ<1|2 - \mu| < 1, so the linearization theorem applies. The dynamics near xx^* is conjugate to the linear map x(2μ)xx \mapsto (2-\mu)x, confirming that orbits converge to xx^* exponentially at rate 2μ|2-\mu|.

At μ=3\mu = 3, we have f3(x)=1f'_3(x^*) = -1, a neutral fixed point where the linearization theorem fails. Indeed, this is precisely where the period-doubling bifurcation occurs, and the behavior transitions from convergence to oscillation.

The linearization theorem bridges local and global dynamics. While it applies only in a neighborhood of a hyperbolic fixed point, understanding local behavior is essential for analyzing bifurcations, constructing invariant manifolds, and developing global portraits of dynamical systems. The proof technique via contraction mappings extends to higher dimensions and forms the foundation for invariant manifold theory.