Gap cleanup followed by minimalization #
Recharting the reduced gap-cleanup output produces a minimal and reduced decomposition. Its weights, mass, gaps, and node count are unchanged by the coordinate change. Increasing the coordinate bounds only weakens the required inverse-power gap threshold.
Minimalize the reduced output of gap cleanup.
Equations
- D.normalized hp C hmod hcenter = EGZ.FlagDecomposition.Rechart.decomposition (D.cleaned hp) C hp hmod hcenter
Instances For
Embed surviving nodes of the normalized pruning into the original node set.
Equations
- D.normalizedNodeEmbedding hp C hmod hcenter = { toFun := fun (x : (D.normalized hp C hmod hcenter).flag.Node) => ↑↑x, inj' := ⋯ }
Instances For
The coordinate change preserves the already established mass-loss bound.
A gap bound in the old coordinates remains valid at larger coordinate bounds after recharting. The denominator uses the old node count.
The subdivision map obtained by cleaning and normalizing the pruned weights.
Equations
- D.normalizedSubdivisionMap hp C hmod hcenter = (D.cleanedSubdivisionMap hp).comp (EGZ.FlagDecomposition.Rechart.subdivisionMap (D.cleaned hp) C hp hmod hcenter)
Instances For
Uniform normalized gap cleanup. The coordinate growth function depends only on dimension, and the prime threshold only on the old uniform bound. The resulting operation is both minimal and reduced.