AFTD Auto-Formalizing Theoretical Domains

pcx_equilibrium_of_root

lemmaverified

Every positive root x of pcx_poly gives a pure-strategy equilibrium of the binary instance with impacts pcx_impact x, in which every type publishes a positive amount at every venue.

Statement

open Finset in
lemma pcx_equilibrium_of_root (x : ℝ) (hx : 0 < x) (hP : pcx_poly x = 0) :
    pcp_is_equilibrium (1 / 2) 2 pcx_theta pcx_mu pcx_cost (pcx_action x) (pcx_impact x) ∧
      ∀ i j, 0 < pcx_action x i j
Source
step towards publication_choice_three_equilibria / publication_choice_three_types_three_equilibria (The Publication Choice Problem, AAAI 2026 (arXiv:2511.13678v2), Theorem 4.1 and Conjecture 4.2)
Verified
06 Oct 2026
Axioms
Classical.choiceQuot.soundpropext
Built on 10
  • pcp_venue_impactVenue impact (eq. (2)): the impact of venue j is the average type of the researchers publishing there, weighted by type densities and…
  • pcp_is_equilibriumPure-strategy equilibrium of a Publication Choice Problem (Definition 1): every researcher type best responds to the venue impacts, and…
  • pcp_half_two_best_responseFor alpha = 1/2 and beta = 2 with positive costs, the closed form of Lemma 3.1, a_j = v_j^4 / (c_j^2 sum_l v_l^4/c_l), solves the…
  • pcx_thetaTypes of the binary instance: theta = (1, 8).
  • pcx_muType densities of the binary instance: (1/3, 2/3).
  • pcx_costCosts of the binary instance: venue 0 costs 1 for both types; venue 1 costs 128 for the low type and 8 for the high type.
  • pcx_assumptionsThe binary instance satisfies the standing assumptions with alpha = 1/2, beta = 2, Assumption 1 (MCR) and Assumption 2 (non-competitive…
  • pcx_polyThe cleared characteristic polynomial of the binary instance in the ratio x of high- to low-type publications at venue 0.
  • pcx_impactVenue impacts of the binary instance as functions of x: ((1+16x)/(1+2x), (1+4096x)/(1+512x)).
  • pcx_actionThe closed-form best-response action profile of the binary instance against the venue impacts pcx_impact x.
Used by 1

Lean source view module on GitHub

open Finset in/-- Every positive root of `pcx_poly` yields a pure-strategy equilibrium of the instance with venue impacts `pcx_impact x`, in which every type publishes a positive amount at every venue. -/lemma pcx_equilibrium_of_root (x : ℝ) (hx : 0 < x) (hP : pcx_poly x = 0) :    pcp_is_equilibrium (1 / 2) 2 pcx_theta pcx_mu pcx_cost (pcx_action x) (pcx_impact x) ∧      ∀ i j, 0 < pcx_action x i j := by  have h1 : (0 : ℝ) < 1 + 2 * x := by linarith  have h2 : (0 : ℝ) < 1 + 512 * x := by linarith  have h3 : (0 : ℝ) < 1 + 16 * x := by linarith  have h4 : (0 : ℝ) < 1 + 4096 * x := by linarith  set V1 := (1 + 16 * x) / (1 + 2 * x) with hV1  set V2 := (1 + 4096 * x) / (1 + 512 * x) with hV2  have hV1p : 0 < V1 := div_pos h3 h1  have hV2p : 0 < V2 := div_pos h4 h2  have hDL : ∑ l, pcx_impact x l ^ 4 / pcx_cost 0 l = V1 ^ 4 + V2 ^ 4 / 128 := by    simp [Fin.sum_univ_two, pcx_impact, pcx_cost, hV1, hV2]  have hDH : ∑ l, pcx_impact x l ^ 4 / pcx_cost 1 l = V1 ^ 4 + V2 ^ 4 / 8 := by    simp [Fin.sum_univ_two, pcx_impact, pcx_cost, hV1, hV2]  have hroot : V1 ^ 4 + V2 ^ 4 / 128 = x * (V1 ^ 4 + V2 ^ 4 / 8) := by    have : x * (V1 ^ 4 + V2 ^ 4 / 8) - (V1 ^ 4 + V2 ^ 4 / 128) =        pcx_poly x / ((1 + 2 * x) ^ 4 * (1 + 512 * x) ^ 4) := by      rw [hV1, hV2, pcx_poly]; field_simp    rw [hP, zero_div] at this; linarith  have hDHp : 0 < V1 ^ 4 + V2 ^ 4 / 8 := by positivity  have hSall : ∀ i, 0 < ∑ l, pcx_impact x l ^ 4 / pcx_cost i l :=    Fin.forall_fin_two.mpr ⟨by rw [hDL]; positivity, by rw [hDH]; positivity⟩  have hpos : ∀ i j, 0 < pcx_action x i j := by    intro i j    simp only [pcx_action]    have hc : ∀ i j, 0 < pcx_cost i j := pcx_assumptions.1.2.2.2.2.2.1    have hI : 0 < pcx_impact x j := by fin_cases j <;> simp [pcx_impact, hV1p, hV2p, ← hV1, ← hV2]    have hS : 0 < ∑ l, pcx_impact x l ^ 4 / pcx_cost i l := hSall i    have := hc i j    positivity  refine ⟨⟨?_, ?_⟩, hpos⟩  · intro i    have hc : ∀ j, 0 < pcx_cost i j := fun j => pcx_assumptions.1.2.2.2.2.2.1 i j    have hS : 0 < ∑ l, pcx_impact x l ^ 4 / pcx_cost i l := hSall i    exact pcp_half_two_best_response (pcx_cost i) (pcx_impact x) hc hS  · intro j    have hDL2 := hDL    have hDH2 := hDH    rw [Fin.sum_univ_two] at hDL2 hDH2    simp only [pcp_venue_impact, pcx_action, Fin.sum_univ_two]    rw [hDL2, hDH2, hroot]    have hx0 : x ≠ 0 := hx.ne'    fin_cases j    · simp [pcx_impact, pcx_cost, pcx_mu, pcx_theta, ← hV1, ← hV2]      field_simp      simp only [hV1, hV2] at *      field_simp      try ring    · simp [pcx_impact, pcx_cost, pcx_mu, pcx_theta, ← hV1, ← hV2]      field_simp      simp only [hV1, hV2] at *      field_simp      try ring
See it in the graphDiscuss this resultChallenge it