Stromquist Preference Sets A(i, j)
For each agent $j \in \mathrm{Fin}\ n$ and piece index $i \in \mathrm{Fin}\ n$, the preference set $A(i, j)$ is the set of simplex points where piece $i$ is a value-maximizer for agent $j$: $$ A(i, j) \;=\; \{x \in S \mid \forall k \in \mathrm{Fin}\ n,\ v(x, j, k) \le v(x, j, i)\}. $$
In Lean: strom_A i j.
Closedness
strom_A_closed: $A(i, j)$ is closed in $\mathrm{Fin}\ n \to \mathbb{R}$.
Proof
The defining condition is a conjunction of $n$ inequalities of the form $v(x, j, k) \le v(x, j, i)$. Each inequality cuts out a closed half-space because both sides are continuous in $x$ ([[social_choice.fair_division.divisible.stromquist_value_continuous]]). The intersection of finitely many closed sets is closed.
Cover property
strom_A_covers: for each fixed $j$, the family $\{A(i, j) : i \in \mathrm{Fin}\ n\}$
covers the simplex $S$.
Proof. For any $x \in S$, agent $j$'s values $v(x, j, k)$ over $k$ form a finite collection of reals; some $i$ achieves the maximum. That $i$ witnesses $x \in A(i, j)$.
Why this matters for KKM
The pair (closed $A(i, j)$, family that covers the simplex) is exactly
the input shape for a KKM-type argument: combined with the "facets-don't-
touch" property of strom_A (strom_A does not intersect the face
opposite vertex $i$), this lets the KKM lemma find a common point in
$\bigcap_i \bigcup_j A(i, j)$ — which yields a fair division.
The unique-preference refinement ([[social_choice.fair_division.divisible.stromquist_unique_preference_sets]]) strengthens the cover to one with pairwise-disjoint witnesses, giving the final EF assignment.
References
- Stromquist, W. (1980). "How to Cut a Cake Fairly". Amer. Math. Monthly 87: 640–644.
- Su, F. E. (1999). "Rental Harmony: Sperner's Lemma in Fair Division". Amer. Math. Monthly 106.