balanced_draft_joint_limit_of_marginals_and_independence
theoremverified
If every agent has the stated marginal Gaussian limit under balanced schedules, and the scaled utilities are jointly independent for every finite number of goods and every deterministic schedule, then the joint Gaussian limit holds under every balanced schedule.
Statement
theorem balanced_draft_joint_limit_of_marginals_and_independence (hmarg : ∀ (n : ℕ), 2 ≤ n → ∀ (schedule : ∀ m, Fin m → Fin n), (∀ m, is_balanced_schedule n m (schedule m)) → ∀ i : Fin n, (∀ m, AEMeasurable (fun v => balanced_draft_scaled_utility n m (schedule m) v i) (balanced_draft_input_law n m : Measure (Fin n → Fin m → ℝ))) ∧ TendstoInDistribution (fun m v => balanced_draft_scaled_utility n m (schedule m) v i) atTop (fun x : ℝ => x) (fun m => (balanced_draft_input_law n m : Measure (Fin n → Fin m → ℝ))) (balanced_draft_gaussian_law n : Measure ℝ)) (hindep : ∀ (n m : ℕ) (s : Fin m → Fin n), iIndepFun (fun i v => balanced_draft_scaled_utility n m s v i) (balanced_draft_input_law n m : Measure (Fin n → Fin m → ℝ))) : ∀ (n : ℕ), 2 ≤ n → ∀ (schedule : ∀ m, Fin m → Fin n), (∀ m, is_balanced_schedule n m (schedule m)) → (∀ m, AEMeasurable (balanced_draft_scaled_utility n m (schedule m)) (balanced_draft_input_law n m : Measure (Fin n → Fin m → ℝ))) ∧ TendstoInDistribution (fun m => balanced_draft_scaled_utility n m (schedule m)) atTop (fun x : Fin n → ℝ => x) (fun m => (balanced_draft_input_law n m : Measure (Fin n → Fin m → ℝ))) (balanced_draft_joint_gaussian_law n : Measure (Fin n → ℝ))
- Source
- Finite-product reduction for the joint Gaussian limit in OP-1493.
- Verified
- 11 Nov 2026
- Axioms
Classical.choiceQuot.soundpropext- Built on 5
- is_balanced_scheduleA picking sequence of m turns among n agents is balanced if each complete block of n turns contains each agent exactly once, and the…
- balanced_draft_scaled_utilityThe vector of normalized sincere-draft utilities, each centered at 2n/(n+1) and multiplied by the square root of the number of goods.
- balanced_draft_input_lawThe finite product probability measure in which every agent-good value is an independent uniform random variable on the unit interval.
- balanced_draft_gaussian_lawThe centered Gaussian probability law whose variance is the nonnegative part of the proposed asymptotic variance; for at least two agents…
- balanced_draft_joint_gaussian_lawThe joint probability law of independent centered Gaussian coordinates with the proposed variance.
Lean source view module on GitHub
open MeasureTheory ProbabilityTheory Filter open scoped Topology theorem balanced_draft_joint_limit_of_marginals_and_independence (hmarg : ∀ (n : ℕ), 2 ≤ n → ∀ (schedule : ∀ m, Fin m → Fin n), (∀ m, is_balanced_schedule n m (schedule m)) → ∀ i : Fin n, (∀ m, AEMeasurable (fun v => balanced_draft_scaled_utility n m (schedule m) v i) (balanced_draft_input_law n m : Measure (Fin n → Fin m → ℝ))) ∧ TendstoInDistribution (fun m v => balanced_draft_scaled_utility n m (schedule m) v i) atTop (fun x : ℝ => x) (fun m => (balanced_draft_input_law n m : Measure (Fin n → Fin m → ℝ))) (balanced_draft_gaussian_law n : Measure ℝ)) (hindep : ∀ (n m : ℕ) (s : Fin m → Fin n), iIndepFun (fun i v => balanced_draft_scaled_utility n m s v i) (balanced_draft_input_law n m : Measure (Fin n → Fin m → ℝ))) : ∀ (n : ℕ), 2 ≤ n → ∀ (schedule : ∀ m, Fin m → Fin n), (∀ m, is_balanced_schedule n m (schedule m)) → (∀ m, AEMeasurable (balanced_draft_scaled_utility n m (schedule m)) (balanced_draft_input_law n m : Measure (Fin n → Fin m → ℝ))) ∧ TendstoInDistribution (fun m => balanced_draft_scaled_utility n m (schedule m)) atTop (fun x : Fin n → ℝ => x) (fun m => (balanced_draft_input_law n m : Measure (Fin n → Fin m → ℝ))) (balanced_draft_joint_gaussian_law n : Measure (Fin n → ℝ)) := by intro n hn schedule hs have hmeas : ∀ m i, AEMeasurable (fun v => balanced_draft_scaled_utility n m (schedule m) v i) (balanced_draft_input_law n m : Measure (Fin n → Fin m → ℝ)) := by intro m i exact (hmarg n hn schedule hs i).1 m have hvec : ∀ m, AEMeasurable (balanced_draft_scaled_utility n m (schedule m)) (balanced_draft_input_law n m : Measure (Fin n → Fin m → ℝ)) := by intro m exact aemeasurable_pi_lambda _ (fun i => hmeas m i) let marginal (m : ℕ) (i : Fin n) : ProbabilityMeasure ℝ := ⟨(balanced_draft_input_law n m : Measure (Fin n → Fin m → ℝ)).map (fun v => balanced_draft_scaled_utility n m (schedule m) v i), Measure.isProbabilityMeasure_map (hmeas m i)⟩ have hm : Tendsto (fun m => fun i => marginal m i) atTop (𝓝 (fun _ : Fin n => balanced_draft_gaussian_law n)) := by apply tendsto_pi_nhds.mpr intro i have hi := (hmarg n hn schedule hs i).2.tendsto have heta : (⟨(balanced_draft_gaussian_law n : Measure ℝ).map (fun x => x), Measure.isProbabilityMeasure_map aemeasurable_id⟩ : ProbabilityMeasure ℝ) = balanced_draft_gaussian_law n := by apply Subtype.ext simp simpa only [marginal, heta] using hi have hp : Tendsto (fun m => ProbabilityMeasure.pi (marginal m)) atTop (𝓝 (balanced_draft_joint_gaussian_law n)) := by have hc := ProbabilityMeasure.continuous_pi.continuousAt.tendsto.comp hm simpa [balanced_draft_joint_gaussian_law, Function.comp_def] using hc have hmap (m : ℕ) : (⟨(balanced_draft_input_law n m : Measure (Fin n → Fin m → ℝ)).map (balanced_draft_scaled_utility n m (schedule m)), Measure.isProbabilityMeasure_map (hvec m)⟩ : ProbabilityMeasure (Fin n → ℝ)) = ProbabilityMeasure.pi (marginal m) := by apply Subtype.ext exact (hindep n m (schedule m)).map_fun_eq_pi_map (fun i => hmeas m i) refine ⟨hvec, ⟨hvec, aemeasurable_id, ?_⟩⟩ have heta : (⟨(balanced_draft_joint_gaussian_law n : Measure (Fin n → ℝ)).map (fun x => x), Measure.isProbabilityMeasure_map aemeasurable_id⟩ : ProbabilityMeasure (Fin n → ℝ)) = balanced_draft_joint_gaussian_law n := by apply Subtype.ext simp simpa only [hmap, heta] using hp