The Informational Curvature Minimization Law states that when a field evolves under curvature-driven smoothing, its total curvature energy decreases monotonically. This remains true even under high-frequency roughness and noisy initial conditions.
Curvature always decreases — sharp bends flatten, smoothness grows.
This law is fundamental in geometric smoothing, informational dynamics, and higher-order diffusion models.
Verified using the uploaded file:
:contentReference[oaicite:3]{index=3}
Curvature operator: lap = roll(A,-1) - 2A + roll(A,1) Curvature energy: R = Σ lap² Evolution: A ← A - α * lap # Laplacian smoothing Stable version: A ← A - α * bi_lap(A) # bi-Laplacian smoothing Result: R(t) strictly decreases.
Both Laplacian and fourth-order bi-Laplacian flows show smooth, monotonic reduction of curvature energy with no rebounds or oscillatory artifacts.
In Fourier modes:
lap(A)ₖ = -k² Aₖ → curvature energy ∼ k⁴ |Aₖ|²
This explains the ultra-fast decay of roughness.
Fourth-order curvature flow improves stability and prevents numerical blow-up, while preserving monotonic decay.
Curvature minimization is the gradient descent of curvature energy, guaranteeing decay.
Minimizing curvature corresponds to reducing bending energy in elastic and informational geometries.
Submissions, analytical proofs, or geometric generalisations may be submitted via:
Submit via Email Submit via Zenodo