Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 13.2: Var[Q]/Var[D]≥1+(2L/m+2L2/m2)(1−ρm)\mathrm{Var}[Q]/\mathrm{Var}[D] \ge 1 + (2L/m + 2L^2/m^2)(1 - \rho^m)Var[Q]/Var[D]≥1+(2L/m+2L2/m2)(1−ρm), with equality when zα=0z_\alpha = 0zα​=0

Proved
SupplyChainTheory.bullwhip_signal_processing

by naimengye · Sep 24, 2026 · Mathlib 0df444a (Lean v4.33.1)

bullwhip-effectforecastingoperations-researchprobabilitysupply-chaintime-series

Theorem 13.2. Demand signal processing creates the bullwhip effect at a single stage.

A retailer faces the stationary AR(1) demand Dt=d+ρDt−1+ϵtD_t = d + \rho D_{t-1} + \epsilon_tDt​=d+ρDt−1​+ϵt​ of Sect. 13.2.2 (d≥0d \ge 0d≥0, −1<ρ<1-1 < \rho < 1−1<ρ<1, ϵt\epsilon_tϵt​ i.i.d. N(0,σ2)N(0,\sigma^2)N(0,σ2)), replenishes with lead time LLL, and, not knowing the demand parameters, sets its base-stock level from a moving average of the previous m≥1m \ge 1m≥1 demands: St=μ^tL+zασ^etLS_t = \hat\mu^L_t + z_\alpha\hat\sigma^L_{et}St​=μ^​tL​+zα​σ^etL​ with μ^tL=L∑i=1mDt−i/m\hat\mu^L_t = L\sum_{i=1}^m D_{t-i}/mμ^​tL​=L∑i=1m​Dt−i​/m and σ^etL=C∑i=1met−i2/m\hat\sigma^L_{et} = C\sqrt{\sum_{i=1}^m e_{t-i}^2/m}σ^etL​=C∑i=1m​et−i2​/m​, ordering Qt=St−St−1+Dt−1Q_t = S_t - S_{t-1} + D_{t-1}Qt​=St​−St−1​+Dt−1​ each period. Then

Var[Qt]Var[Dt]  ≥  1+(2Lm+2L2m2)(1−ρm),\frac{\mathrm{Var}[Q_t]}{\mathrm{Var}[D_t]} \;\ge\; 1 + \Big(\frac{2L}{m} + \frac{2L^2}{m^2}\Big)\big(1 - \rho^m\big),Var[Dt​]Var[Qt​]​≥1+(m2L​+m22L2​)(1−ρm),

and the bound is tight when zα=0z_\alpha = 0zα​=0: with no safety stock the ratio equals the right-hand side.

The bound exceeds 111 whenever L>0L > 0L>0, so a positive lead time together with forecasting is enough to make the orders more variable than the demands, even when ρ=0\rho = 0ρ=0 and the demands are independent. The book draws out the comparative statics: the bound decreases in mmm (smoother forecasts), increases in LLL, decreases in ρ\rhoρ for ρ≥0\rho \ge 0ρ≥0, and for ρ<0\rho < 0ρ<0 behaves differently for odd and even mmm. Theorems 13.6 and 13.7 apply the same bound stage by stage to a serial supply chain.

Formalization Note The proof runs through the order identity, the autocovariance (13.4), the vanishing of the cross term (Lemma 13.1) and the variance of the demand part, all separate items; the ratio is stated with Var[Dt]=σ2/(1−ρ2)>0\mathrm{Var}[D_t] = \sigma^2/(1-\rho^2) > 0Var[Dt​]=σ2/(1−ρ2)>0. The book's constant CLρC_{L\rho}CLρ​ in the estimate of σ^\hat\sigmaσ^ is a free parameter, which the bound does not involve.

Preamble
import Definitions.Def_SupplyChainTheory_bullwhip
Formal statement
namespace SupplyChainTheory

theorem bullwhip_signal_processing {Ω : Type*} [MeasurableSpace Ω] {P : MeasureTheory.Measure Ω}
    [MeasureTheory.IsProbabilityMeasure P] (X : AR1Demand P) (C z : ℝ)
    (L m : ℕ) (hm : 0 < m) (t : ℤ) :
    ProbabilityTheory.variance (X.order C z L m t) P / ProbabilityTheory.variance (X.D t) P
        ≥ 1 + (2 * (L : ℝ) / m + 2 * (L : ℝ) ^ 2 / (m : ℝ) ^ 2) * (1 - X.rho ^ m)
      ∧ (z = 0 →
          ProbabilityTheory.variance (X.order C z L m t) P / ProbabilityTheory.variance (X.D t) P
            = 1 + (2 * (L : ℝ) / m + 2 * (L : ℝ) ^ 2 / (m : ℝ) ^ 2) * (1 - X.rho ^ m)) := by sorry

end SupplyChainTheory
Source
Lawrence V. Snyder and Zuo-Jun Max Shen, Fundamentals of Supply Chain Theory, 2nd ed., Wiley 2019, DOI 10.1002/9781119584445, p. 545, Sect. 13.2.2, Theorem 13.2, Eq. (13.13): 'The bound is tight when z_alpha = 0'; the model is pp. 542-544, Eq. (13.1)-(13.9); after Chen et al. (2000)
Human review
  • Endorsed by Shuze Chen · Sep 25, 2026

    Confirmed by the moderator at approval.

  • Endorsed by naimengye · Sep 25, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me