Claim 5.2, proof — the bonus is maximized by executing in minimal time
ProvedAlgMechDesign.CompBonus.bonus_max_minimal_execmechanism-designp2o-batch-p200bp2o-gran-per-chapterp2o-plan-paperp2o-v1scheduling
Fix an allocation , an agent of type , and declarations of the other agents; write for the resulting profile. Let be any vector of actual times in which agent performs each of its tasks no faster than possible, for (the other coordinates are arbitrary, since they are not used). Then
and is exactly the corrected time vector of agent when it performs its tasks in minimal time .
In words: for every allocation, agent 's bonus is maximized when it executes its assignments in minimal time. This is the second observation of the proof of Claim 5.2.
Preamble
import Mathlib import Definitions.Def_AlgMechDesign_CompBonus_Model import Definitions.Def_AlgMechDesign_CompBonus_Mechanism
Formal statement
namespace AlgMechDesign.CompBonus
/-- Proof of Claim 5.2 (p. 188): for every allocation `x`, the bonus of agent `i`, whose true type
is `t i` and whose fellow agents declared `t l`, is maximized when `i` performs its tasks in
minimal time: for every vector of actual times `tt` with `t i j ≤ tt j` on the tasks of `i`,
`-g(x, corrⁱ(x, t, tt)) ≤ -g(x, corr*(x, t))`, and `corr*(x, t)` is `corrⁱ` at minimal times. -/
theorem bonus_max_minimal_exec {n k : ℕ} [NeZero n] (x : Fin k → Fin n)
(t : Fin n → Fin k → ℝ) (i : Fin n) (tt : Fin k → ℝ)
(htt : ∀ j, x j = i → t i j ≤ tt j) :
-gT x (corr i x t tt) ≤ -gT x (corrStar x t) := by sorry
end AlgMechDesign.CompBonus
Source
Nisan, Ronen, Algorithmic Mechanism Design, Games Econ. Behav. 35, 2001, p. 188, proof of Claim 5.2, first paragraph, third sentence ("… and that for every allocation x the bonus for an agent i is maximized when executing its assignments in minimal time")
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.