Winding Number: Crossing Analysis #
Technical analysis of curve crossings for generalized winding numbers. Contains the core monotonicity, cutoff boundary, and direction convergence lemmas.
Main Results #
piecewiseC1Immersion_norm_strictMono_near_crossing— local monotonicity at crossingsexists_cutoff_boundary_times— existence of cutoff boundary timesexp_cutoff_integral_eq_ratio— exponential of cutoff integral equals direction ratiocrossing_ratio_tendsto— direction convergence as ε → 0tendsto_exp_cutoff_integral_crossing— exp(R(ε)) → exp(-iα)
Helper: g = ‖γ(·) - z₀‖ is strictly decreasing on a left neighborhood of t₀ and strictly increasing on a right neighborhood, when γ is an immersion at t₀. This is the key "local monotonicity" fact that makes the cutoff boundary well-defined.
Extended version of exists_cutoff_boundary_times that also exposes the
strict monotonicity interval and the bounds δ ≤ ‖γ(l) - z₀‖,
δ ≤ ‖γ(r) - z₀‖.
For a closed piecewise C¹ immersion, when the cutoff integral is split at boundary times where ‖γ-z₀‖ = ε with strict inequality outside, the exponential equals the ratio (γ(σ₁)-z₀)/(γ(σ₂)-z₀) by FTC + closedness.
Direction convergence: as ε → 0, the ratio (γ(σ₁(ε))-z₀)/(γ(σ₂(ε))-z₀)
(where σ₁(ε), σ₂(ε) are the boundary times from exists_cutoff_boundary_times)
converges to exp(-i·angleAtCrossing). This follows from the immersion property:
γ(σ₁)-z₀ ≈ L_left·(σ₁-t₀) with σ₁-t₀ < 0, so direction → -L_left/|L_left|,
and γ(σ₂)-z₀ ≈ L_right·(σ₂-t₀) with σ₂-t₀ > 0,
so direction → L_right/|L_right|.
The ratio of directions is exp(-i·α) where α = arg(L_right) - arg(-L_left).
Core analysis: exp(R(ε)) → exp(-iα) as ε → 0, where R(ε) is the
cutoff integral ∫ 1_{‖γ-z₀‖>ε} (γ-z₀)⁻¹ γ' and α is the crossing angle.
Proof strategy (H-W Proposition 2.2):
- For each small
ε, the set{t : ‖γ(t)-z₀‖ ≤ ε}is a single interval(σ₁(ε), σ₂(ε))containingt₀(by continuity + isolated crossing). - By piecewise FTC on segments where
‖γ-z₀‖ > ε(using the G-function technique fromexp_integral_eq_endpoint_ratio_piecewise):exp(R(ε)) = (γ(σ₁)-z₀)/(γ(a)-z₀) · (γ(b)-z₀)/(γ(σ₂)-z₀). - By closedness
γ(a) = γ(b):exp(R(ε)) = (γ(σ₁)-z₀)/(γ(σ₂)-z₀). - Since
‖γ(σ₁)-z₀‖ = ε = ‖γ(σ₂)-z₀‖, this ratio has modulus 1. - By the immersion property: as
ε → 0,σ₁ → t₀⁻andσ₂ → t₀⁺, andarg(γ(σ₁)-z₀) → arg(-L_left),arg(γ(σ₂)-z₀) → arg(L_right). - Therefore
exp(R(ε)) → exp(i(arg(-L_left) - arg(L_right))) = exp(-iα).