Problems / combinatorics
combinatorics / Extremal set theory
The Daykin–Frankl conjecture on convex subsets of the Boolean lattice
Let P⊆Qn be convex. Williams proves the stronger statement that for every k≥0,
w(P×Qk)≥w(Qn+k)∣P∣2−n,
where w denotes poset width.
Taking k=0 gives
w(P)≥∣P∣(⌊n/2⌋n)2−n,
which is exactly the Daykin-Frankl conjecture.
The proof proceeds by induction on n, reducing the step to a structural lemma for a convex subset of R×Q1 and carefully recombining antichains from its two layers.