Research notes on every problem, and a library of the papers behind them. Built from the open erdos repository.
Updated
Source. Theorem 7, p. 9, with its proof on pp. 9–10, of David Ellis, Irredundant families of subcubes, arXiv:1003.2960v1 (2010), published in Mathematical Proceedings of the Cambridge Philosophical Society 150(2) (2011), 257–272, as identified on the source card. Labels and pages are those of arXiv:1003.2960v1.
Statement
The Hamming ball of centre and radius is (p. 2); subcubes and irredundance are as on the Theorem 4 page.
Theorem 7 (p. 9). Let be a Hamming ball of radius in . If is an irredundant family of -subcubes of , each with a private vertex in , then .
Equality holds when is the family of all -subcubes through the centre of (p. 10). Fixing private vertices and averaging over all Hamming balls of radius recovers Theorem 4 (p. 10). The introduction (p. 3) gives Theorem 7 in the form: if one private vertex is chosen for each member of an irredundant family, any Hamming ball of radius contains at most of them.
Proof pointer
Pages 9–10. Take . For each member , with private vertex in , let be the sub-subcube of between and the top vertex of . The claim (7), p. 9, bounds a weighted count of the members whose contains a given -set , by Theorem 3. Summing (7) over the vertices of layer , each member contributes exactly .
Dependencies
Read depth. Claims checked: the statement on p. 9 and the equality and averaging remarks on p. 10 were read clause by clause. The proof was read but not checked step by step.
Bears on
No Erdős problem.