Value-Separated Vertices are Rigid — the Telescoping Argument #
Discovery B of the symmetry-breaking programme
(research/attacks/symmetry-breaking.md): call a vertex v
value-separated if no non-neighbour of v shares a card-degree with a
neighbour of v. Then v is rigid, hence any graph possessing such a
vertex is reconstructible — a strict extension of the regular case (where
every card-degree class is trivially all-neighbours or all-non-neighbours).
The proof is the telescoping argument: for each card-degree value t, the
counts of neighbours (nbrCount), non-neighbours (nonNbrCount), the full
class (classCount), and the G-degree classes (fullDegCount) satisfy
classCount t = nbrCount t + nonNbrCount t(partition);fullDegCount t = shiftNbrCount t + nonNbrCount twhereshiftNbrCount (t+1) = nbrCount t,shiftNbrCount 0 = 0(the deficit stamp, pointwise);
and the deck matches classCount (via the card isomorphism) and
fullDegCount (via the degree multiset and the deleted degree) across the
two graphs. Induction on t then forces the neighbour counts to agree
classwise, and value-separation upgrades classwise agreement to the
pointwise link condition.
Main results #
SimpleGraph.ValueSeparated— the separation predicate;SimpleGraph.ValueSeparated.rigidVertex— value-separated vertices are rigid;SimpleGraph.nonempty_iso_of_valueSeparated— graphs with a value-separated vertex are reconstructible.
v is value-separated if no non-neighbour shares a card-degree
value with a neighbour: every card-degree class of G − v is
all-neighbours or all-non-neighbours of v.
Equations
- G.ValueSeparated v = ∀ (x y : { w : V // w ≠ v }), ¬G.Adj v ↑x → G.Adj v ↑y → (G.deleteVert v).degree x ≠ (G.deleteVert v).degree y
Instances For
Value-class counts on a card #
Number of neighbours of the deleted vertex with card-degree t.
Equations
Instances For
Number of non-neighbours of the deleted vertex with card-degree t.
Equations
Instances For
Transfer across a degree-matched card isomorphism #
Value-separated vertices are rigid #
Value-separated vertices are rigid (Discovery B): classwise neighbour counts are forced by the telescoping recurrence, and value-separation makes classwise agreement pointwise — the identity correction works.
Discovery A, public deck form: the value-class neighbour profile of
every matched card is deck-forced. For same-deck graphs there is a vertex
matching σ such that, for every vertex v and every card-degree value
t, the number of v-neighbours of card-degree t in G − v equals the
number of σv-neighbours of card-degree t in H − σv (and likewise for
non-neighbours). In particular the neighbour-degree multiset of v is
deck-determined along the matching: the only labelling freedom the deck
ever leaves is which vertices within a card-degree class are the
neighbours. This is the formal content of the shift-recurrence discovery in
research/attacks/symmetry-breaking.md.
Graphs with a value-separated vertex are reconstructible — Discovery
B combined with the rigid-card criterion. Strictly extends
nonempty_iso_of_regular.
A card with pairwise-distinct degrees has no mixed class at all, so its deleted vertex is value-separated; a graph with an injective-degree card is reconstructible. A purely card-side recognizable criterion.
Every vertex of a regular graph is value-separated: non-neighbours keep
card-degree d, neighbours drop to d − 1.