Proof of Theorem 9, last sentence (first half) — (1/N)·d_i(M) ≥ MmS^N_{d_i}(M)
ProvedMmSChores.RoundRobin.mms_le_averagefair-divisionmaximin-sharep2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-strengthenedp2o-v1
Let agents have additive utilities over a finite set of items, and let
be agent 's max-min share, the maximum ranging over all allocations of into (possibly empty) bundles. Then for every agent ,
The max-min share never exceeds the proportional share. The paper states this "in every chores setting"; it holds for arbitrary real additive utilities.
Formalization Note The max-min share is the published LittleCharity.MMS.mms (bval d i) N univ, a supremum over labelled maps Fin m → Fin N (bundles may be empty) of the minimum bundle value; for N ≥ 1 both index types are finite and nonempty, so it is the true max-min.
Differences from the source:
- generalization: in every chores setting (d_i(j) ≤ 0, d_i ≢ 0) → arbitrary real utilities d_i, no sign condition; reason: the inequality holds for every additive utility and specializes to chores instances
Preamble
import Mathlib import Definitions.Def_LittleCharity_MMS_Setting import Definitions.Def_MmSChores_RoundRobin_Setting
Formal statement
namespace MmSChores.RoundRobin
open Finset
/-- Proof of Theorem 9 (PDF p. 4), first half of the last sentence: `1/N · d_i(M) ≥ MmS^N_{d_i}(M)`.
It holds for every real additive utility, with no sign condition. -/
theorem mms_le_average {N m : ℕ} (hN : 0 < N) (d : Fin N → Fin m → ℝ) (i : Fin N) :
LittleCharity.MMS.mms (bval d i) N univ ≤ (1 / (N : ℝ)) * bval d i univ := by sorry
end MmSChores.RoundRobin
Source
Aziz, Rauchecker, Schryen & Walsh, Algorithms for Max-Min Share Fair Allocation of Indivisible Chores, AAAI 2017, PDF p. 4, proof of Theorem 9, sentence after the display after eq. (1)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.