The linear strategies of Example 1 solve the linked differential equations (3a)–(3b)
ProvedChatterjeeSamuelson.UniformEquilibrium.linear_solves_linked_odesbargaininggame-theoryincomplete-informationp2o-batch-p200ap2o-gran-per-chapterp2o-plan-paperp2o-v1
Let and , let both beliefs be uniform on , so on with density , and let and . Then:
- (3a) for every seller value and every with ,
- (3b) for every buyer value and every with ,
Here plays the role of the paper's and that of . Equations (3a)–(3b) are the linked differential equations that Theorem 2 of the paper derives as necessary conditions for well-behaved equilibria; this statement records that the linear strategies of Example 1(a) satisfy them for uniform beliefs, which is the paper's justification of Example 1(a).
Formalization Note and are the distribution function of the uniform belief ; the density is written as a constant; derivatives are Mathlib's deriv. The inverse functions are replaced by bound variables with , which determine them uniquely.
Preamble
import Mathlib import Definitions.Def_ChatterjeeSamuelson_Shared_unif import Definitions.Def_ChatterjeeSamuelson_Shared_sellerLinear import Definitions.Def_ChatterjeeSamuelson_Shared_buyerLinear open MeasureTheory ProbabilityTheory Set
Formal statement
namespace ChatterjeeSamuelson.UniformEquilibrium
/-- The linear strategies of Example 1(a) solve the linked differential equations (3a) and
(3b) (Chatterjee & Samuelson, *Bargaining under Incomplete Information*, Oper. Res. 31(5)
1983, §3, pp. 842–843 [PDF 8–9], unnumbered text after Example 1: "By simple inspection
one can check that linear offer strategies above constitute solutions of Equations 3a and
3b when both distribution functions are linear."; equations (3a), (3b) are Theorem 2,
p. 840 [PDF 6]).
With `F_b = F_s = cdf (unif v̄)` (equal to `v/v̄` on `[0, v̄]`) and densities
`f_b = f_s = 1/v̄`, `Slin = sellerLinear k v̄`, `Blin = buyerLinear k v̄`:
* (3a) `k F_b(y) S′(y) + f_b(y) S(y) = B⁻¹(S(y)) f_b(y)`: for every seller value
`y ∈ [0, v̄]` and every `x` with `Blin x = Slin y`,
`k · F_b(y) · Slin′(y) + (1/v̄) · Slin(y) = x · (1/v̄)`;
* (3b) `(1 − k)(1 − F_s(x)) B′(x) − f_s(x) B(x) = −S⁻¹(B(x)) f_s(x)`: for every buyer value
`x ∈ [0, v̄]` and every `y` with `Slin y = Blin x`,
`(1 − k)(1 − F_s(x)) · Blin′(x) − (1/v̄) · Blin(x) = −(y · (1/v̄))`.
*Formalization Note.* `B⁻¹(S(y))` is written as a bound variable `x` with `Blin x = Slin y`
(both linear maps are injective for `0 ≤ k ≤ 1`, so `x` is unique), and `S⁻¹(B(x))` likewise.
`S′`, `B′` are Mathlib's `deriv`. The density `1/v̄` is written as a constant: it is the
derivative of `F(v) = v/v̄` on `(0, v̄)`. The identities are checked on the whole value
interval `[0, v̄]`, including values where the equilibrium strategy of Example 1(a) leaves
the linear branch. -/
theorem linear_solves_linked_odes (k vbar : ℝ) (hk0 : 0 ≤ k) (hk1 : k ≤ 1) (hvbar : 0 < vbar) :
(∀ y ∈ Icc (0 : ℝ) vbar, ∀ x : ℝ, Shared.buyerLinear k vbar x = Shared.sellerLinear k vbar y →
k * cdf (Shared.unif vbar) y * deriv (Shared.sellerLinear k vbar) y
+ (1 / vbar) * Shared.sellerLinear k vbar y = x * (1 / vbar)) ∧
(∀ x ∈ Icc (0 : ℝ) vbar, ∀ y : ℝ, Shared.sellerLinear k vbar y = Shared.buyerLinear k vbar x →
(1 - k) * (1 - cdf (Shared.unif vbar) x) * deriv (Shared.buyerLinear k vbar) x
- (1 / vbar) * Shared.buyerLinear k vbar x = -(y * (1 / vbar))) := by sorry
end ChatterjeeSamuelson.UniformEquilibrium
Source
Chatterjee & Samuelson, Bargaining under Incomplete Information, Oper. Res. 31(5) (1983), https://doi.org/10.1287/opre.31.5.835, pp. 842–843 [PDF 8–9], text following Example 1 ("By simple inspection one can check…"); equations (3a), (3b) from Theorem 2, p. 840 [PDF 6]
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.