The hub-cross firing law and the big-clean-twin structure #
The spectral closer for the hub–hub mediation channel of the m-bound, together with the rigidity of the twin population it controls.
Main results #
algConn_le_two_of_hub_cross_pair— the hub-cross firing law: two non-adjacent degree-3 verticesu ≠ vwith no common neighbour, partners all of degree≤ 4(hub degreesD, D' ≥ 4unbounded), and whose only possibleN(u)–N(v)edge is the hub–hub edgeg ~ g', forcealgConn G ≤ 2. It strictly extends the cross-free usable far-pair law: instantiating the weighted double-star master withau = av = 1, hub weightss = 1/(D−1)and a balance-shifted partner weight makes the hub leak-identity(1−s)² + (D−2)s² = 1 − syield worst-case slack−(s+s')(the hub–hub cross is spectrally cheap).usable_twin_partners_deg_le_four— σ-rigidity: a usable degree-3 twin of a degree-≥ 9hub has both other partners of degree≤ 4(σ(9) + σ(5) + σ(4) > 2). So the firing law applies to every cross-cloud twin pair mediated only by a hub–hub edge.bigCleanTwinsand its structure lemmasbigCleanTwins_nbr_deg(neighbours have degree4or≥ 9),bigCleanTwins_unique_big(exactly one big neighbour, two degree-4 partners),sum_big_inc_le,sum_small_inc_le— the bipartite incidence counts feeding the hub-cross coverage argument.
σ-rigidity of big-hub twins. A usable degree-3 vertex t (sigS ≤ 2)
adjacent to a hub g of degree ≥ 9 has both partners of degree ≤ 4,
provided neither partner is degree-3 (i.e. off the M-cluster):
σ(9) + σ(5) + σ(4) = 6/7 + 2/3 + 1/2 = 85/42 > 2.
The hub-cross firing law. u ≠ v degree-3, non-adjacent, no common
neighbour; g ∈ N(u), g' ∈ N(v) hubs of degree ≥ 4 (unbounded above); all
other neighbours (partners) of u and of v have degree ≤ 4; and every
N(u)–N(v) edge is the hub–hub edge (g, g'). Then algConn G ≤ 2 — the
single hub–hub cross is spectrally cheap and cannot block the double-star
certificate. Weights: au = av = 1, hub weights 1/(deg−1), u-partners
½, v-partners ½ + (s−s')/2; worst-case quadratic slack = −(s+s') (edge
present) resp. −(s+s')−2ss' (absent).
Big clean twins #
The big clean twins bigCleanTwins G = {u : deg u = 3 ∧ sigS u ≤ 2 ∧ (∃ x ~ u, deg x ≥ 9) ∧ (∀ x ~ u, deg x ≠ 3)} are the usable degree-3 vertices
with a big neighbour and no degree-3 neighbour. By σ-rigidity their
neighbourhoods are completely determined (bigCleanTwins_nbr_deg,
bigCleanTwins_unique_big), and the bipartite incidence counts sum_big_inc_le
/ sum_small_inc_le feed the hub-cross pair-coverage count.
The big clean twins: usable degree-3 vertices with a degree-≥ 9
neighbour and no degree-3 neighbour.
Equations
- ACMax.bigCleanTwins G = {u : Fin n | G.degree u = 3 ∧ ACMax.sigS G u ≤ 2 ∧ (∃ x ∈ G.neighborFinset u, 9 ≤ G.degree x) ∧ ∀ x ∈ G.neighborFinset u, G.degree x ≠ 3}
Instances For
Neighbourhood rigidity. Every neighbour of a big clean twin has degree
4 or degree ≥ 9.
Unique big neighbour. A big clean twin has exactly one big neighbour:
there is a big neighbour g, and every other neighbour has degree ≤ 4.
Bipartite incidence exchange (indicator double count):
Σ_{w ∈ s} |N(w) ∩ t| = Σ_{u ∈ t} |N(u) ∩ s|.
Big-incidence exchange. The incidences between big vertices and big
clean twins number at most |bigCleanTwins| — each twin has exactly one big
neighbour.
Small-incidence exchange. The incidences between degree-≤ 8 vertices
and big clean twins number at most 2·|bigCleanTwins| — each twin has exactly
two degree-4 partners.
The M-population under the dichotomy: at most 2 degree-3 vertices
have a degree-3 neighbour.